You are on page 1of 210

Dianwei Qian · Jianqiang Yi

Hierarchical
Sliding Mode
Control for Under-
actuated Cranes
Design, Analysis and Simulation

Hierarchical Sliding Mode Control
for Under-actuated Cranes

Dianwei Qian Jianqiang Yi

Hierarchical Sliding
Mode Control
for Under-actuated Cranes
Design, Analysis and Simulation

123

Dianwei Qian Jianqiang Yi
School of Control and Computer Institute of Automation
Engineering Chinese Academy of Sciences
North China Electric Power University Beijing
Beijing China
China

Additional material to this book can be downloaded from http://extras.springer.com.

ISBN 978-3-662-48415-9 ISBN 978-3-662-48417-3 (eBook)
DOI 10.1007/978-3-662-48417-3

Library of Congress Control Number: 2015949449

Springer Heidelberg New York Dordrecht London
© Springer-Verlag Berlin Heidelberg 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.

Printed on acid-free paper

Springer-Verlag GmbH Berlin Heidelberg is part of Springer Science+Business Media
(www.springer.com)

Dedicated to my wife Weiwei Zhao and our daughter Siqi Qian as well as to my parents Yubao Qian and Lixiang Bai with love and gratitude Dianwei Qian Dedicated to my wife Li Zhang and our children with love and gratitude Jianqiang Yi .

shipyards and construction sites. sliding motion can be depicted by phase plane and this appears to be beneficial both for scientific researches and studies. Accordingly. disaster sites. most of the current researches only focus on some special control applications. such as warehouses. Among the diversity of control methods. With hierarchical sliding mode control architectures. Capturing the structure characteristic of overhead cranes. The key feature of hierarchical sliding mode control architectures is the hier- archical sliding surfaces with the guaranteed stability. some appropriate control structures are constructed and hierarchical sliding mode control is developed. rather than a systematic methodology. This book provides readers with a comprehensive overview of sliding mode control for overhead crane systems with detailed proofs of the fundamental results. With the under- standing that the physical structure of overhead crane systems is dismantled. By explicitly building the structure specification into the problem formulation. various control methods have been presented. In spite of these remarkable advances in this field. Concerning the control problem. it is possible to construct and analyze novel sliding surface structures. nuclear plants. The methodology of hierarchical sliding mode control fills the gap between sliding mode control and its applications in overhead cranes. Overhead crane control has been paid more and more attention in recent years. this is possible to achieve novel control structures by the method of sliding mode control. It is a systematic and effective design tool. vii . sliding mode control is recognized as one of the most efficient design tools. Hierarchical sliding mode control presents several control structures for the sliding mode control applications in overhead cranes.Preface Overhead cranes are widely used in many places. numerous theoretical studies and practical implementations have been carried out. Many control designs and applications in the fields of sliding mode control for overhead cranes have been published in various journals and conference proceed- ings. which has both theoretical and practical significances.

MATLAB codes about the simulations are also presented. please feel free to contact Dianwei Qian by E-mail: dianwei. Chapter 5 extends the method of hierarchical sliding mode control to accom- modate unmatched uncertainties.qian@gmail. states the open problems. Chapter 1 starts with a brief introduction of overhead crane systems. If you have questions about algorithms and simulation programs.com. MATLAB codes about the simulations are also available in the appendix part. The compensator and the controller work together to realize the robust overhead crane control. which can be run successfully in MATLAB 7.5. Numerical simulations illustrate the feasibility of these designed hier- archical structures. Then. The book can be used for teaching a graduate-level special-topics course in sliding mode control. Beijing Dianwei Qian Jianqiang Yi . The chapter proceeds with analysis of oscillations for pendulum-type motions according to the linearized models of the two types of overhead cranes. it is proved that both the hierarchical sliding sur- faces and the whole control system are of asymptotically stability in the sense of Lyapunov. Some simulations are conducted to verify the effectiveness of the presented control scheme. equations of motion for single-pendulum-type overhead cranes are presented. this chapter proceeds with the design of intelligent compensator based on the hierarchical structure. Some simulation results are demonstrated. Uncertainties of the two models are considered next. First. dynamics of double-pendulum-type overhead cranes are described. This chapter proceeds with analysis of some typical control problems associated with sliding mode control for overhead cranes. Since the compensator design needs a strict assumption. For each hierarchical structure. In this book.342 version or in other more advanced versions. It proceeds with a brief historical overview of sliding mode control. MATLAB codes about the simulations are also attached in the appendix part. It starts with compensator design for the unmat- ched uncertainties of overhead cranes.viii Preface The book has six chapters. Chapter 3 introduces several typical design methods of sliding mode control. Chapter 4 presents hierarchical sliding mode control for overhead cranes. Each chapter concludes with appendices about simulation programs. The chapter proceeds with applications of these design methods for overhead cranes. Chapter 2 investigates modeling of overhead crane systems.0. A review about overhead crane control is considered next. all the control algorithms and their programs are described sepa- rately and classified by the chapter name. Three hierarchical structures are designed for single-pendulum-type overhead cranes and one hierarchical structure is considered for double-pendulum-type overhead cranes. Chapter 6 summarizes some of the further extensions not captured within this book. and the challenges for future thinking.

dissertations and technical papers contribute to the chapters in this book. for his help and support with crane control applications.Acknowledgments We would like to acknowledge a number of research scientists. we are especially thankful to Professor Dongbin Zhao of the Institute of Automation. We are grateful to Dr. and Hong Liu from the School of Control and Computer Engineering. Chinese Academy of Sciences. Diantong Liu. in the initial stages of the theory’s development. and Dr. con- tributed to the development of the results in this book while working on their dissertations.D. At last but not least. Wei Wang. love. postdoctoral fellows. and support. Among these. we would like to thank our families for their unconditional dedication. Dr. Beijing Dianwei Qian Jianqiang Yi ix . Graduate students Jinrong Guo. BinbinYang. Chinese Academy of Sciences. and graduate students who have worked with us at different times and whose Ph. Yinxing Hao for numerous useful discussions on the design and the analysis problems of the methods in this book during their stay at the Institute of Automation. and to whom—with our humble gratitude—we dedicate this book. North China Electric Power University.

. . . . . . . . . . . .7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20 . . 38 H Simulink Model to Plot Figs.1 Theoretical Challenges . . . . . .3. . . 36 E Simulink Model to Plot Figs. .11 . . . . . . .1 Open-Loop Control . . . . . . . . . . . . . . . . . . . . . . . . . . 41 References .5a. . . . . . . . . . . . . . . . . . . . .3 Tower Cranes . . . . . . . . . . . . .1. . .2. . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . . 5 1. . . . .16. . 35 C Matlab Codes to Plot Fig. . . . . . . . . . . . . . . . . . .5b. . . .2. . . . . . . . . 1. . . .7 Sliding Order and Sliding Sets . . . 1. . . . . .1 Crane Types . . . . 17 1. . .1. . . . . . . . . . . . . . . . 1. . . . . . . . . 25 1.12 . . . . 23 1. . 1. . . . 1 1. . . . . . . . . . . 1. . . . . . . . .8 and 1. . . . . . . . . . . . . . . . . .3 Equivalent-Control-Based Sliding Mode Control . .3 A Review of Crane Control. . . . . . . . . . . . . . 33 1. 1 1.2. . . . . . 35 A Matlab Codes to Plot Fig. . . . .2 Practical Challenges . . . . . . . . . . .4. .2 Sliding Mode Control. . . . . . . . . . . . . . . 2 1. . . . . . . . . . . . . . . 1. . . . . . . . . . . . . . . . . . . . . . . . . .5 Sliding Mode Control Design for State Space Model .1. . 38 G Simulink Model to Plot Figs. . . . . . . . . . 1. . . . . 11 1. . . . . . . . . .2 Closed-Loop Control . . . . 34 Appendices . . . . . . . . . . . . . . . . . . . . . 35 D Simulink Model to Plot Figs. . . . . . . . . .6 Robustness Against Uncertainties. . . . . .1 Overhead Cranes . . . . . . . .4. . . . . . . .4 Chattering Reduction . 1. . . .2. . . . . . . . . 35 B Matlab Codes to Plot Fig. . . . .13 and 1. . . . . . . 37 F Matlab Codes to Plot Fig. 13 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 and 1. . . . . . 39 I Simulink Model to Plot Figs. . . . . . . . . . . . . . . . . . . . . . 42 xi . . . .14. . . . . . . . .1 Variable Structure Control . . . . . . . . .2. . . . . . . 33 1. . . . 5 1. . . . 23 1. 3 1. . . . . .10 and 1. . . 22 1. . . . . . .17.9 . . . . . . . . . .4 Challenges of Sliding Mode-Based Crane Control . . . . . . . .2 Boom Cranes .Contents 1 Introduction. . . . . . . . . . . 8 1. . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . 15 1. . . . . . . .2. . . . . . .15 and 1. . . . . . . . . . . . . . . . . . . . . 2 1. .2 Review of Sliding Mode Control . . . . . . . .18. . . . . . . . . 1. . 1. . . . 1. . . . . . . .

.4 Terminal Sliding Mode Control . . . .4. . . . . . . . . .. . . . . . . . . .. . . . . . . . . . . . . . . . . . .1. . . . . . . . . .1. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .5. .. . . .. . . . . . .2 Stability Analysis. . . . . . . . . .1. . . . . . . . . . . . . . 51 2. . . . . . . . . . 79 3. . . . . . . .3 Simulations of Single-Pendulum-Type Cranes by TSM. . . . . . .. . . . . . . . . . . . . .. . . . . . . . . . . . . . 88 3. . . 51 2. . . . . . . . . . . . . . . . . . . . 80 3. . ... . . . . . . . . . . . . . . .. . . . . . . .. . . . . 93 3. . . . .4 Simulations of Uncertain Single-Pendulum-Type Cranes by TSM . . . . . . . . . . . 57 2. 66 3 Overhead Crane Control by Sliding Mode Methods. .1 Control Design . . . . . 69 3. .3 Linearized Model. .3. . . . . . . . .. . .. . . . . . . . . . . . . 68 3..4 Simulations of Uncertain Single-Pendulum-Type Overhead Cranes . . . .. . . . . . . .. . . . .2. . . 79 3. . .6 Linearized Model.. . . . . .4. . . . .. . . . . . . . . .1. . . . . . . . . . .. . . . . . . . . . 65 A Matlab Codes to Plot Fig. . . . . . . . . . . . . . . . . . . 66 References . . . .. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 . . . . .3. . .5. .3. . . . 63 2. .3. . . .1 Control Design of Single-Pendulum-Type Overhead Cranes . . . . 84 3. . . . . . . . . . . . . . . . . . . .. . . . . . . .. .1 Problem Description . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . .5 Model with Uncertainties . . . . . . . . . . . 56 2. . . . . .2. . . . . . . . . . . .. .. . . .4 Modeling of Double-Pendulum-Type Cranes . . . . . . . . . . 82 3.. . . .. 68 3. . . . . . . . . . . . . . . .3 Integral Sliding Mode Control . . . . .. . . . . . 51 2. . . . . . .1. . . . . . . .5 Extensions of Double-Pendulum-Type Overhead Cranes .. . 55 2. . .2 First-Order Sliding Mode Control. . . . 71 3. . . . .3 Simulations of Single-Pendulum-Type Cranes by ISMC . . . . . .. .1 Modeling of Single-Pendulum-Type Cranes. . . . . . . . . 94 .2 Stability Analysis of the Single-Pendulum-Type Crane Control System . . 67 3.3 Simulations of Nominal Single-Pendulum-Type Overhead Cranes . . . . . . 73 3. . . . . .. .2 Stability Analysis. . .. . . .2. . . .. . . . . . . . . . . . . .. . . . . . . .2 Model with Uncertainties .. . . . . . . . . . 2.. 91 3. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . . . . . .. . . 67 3. . . . .. . . . . . . . . .2. . . . . . . . .. . . . . . . . .1 Modeling . . . . .. . .. . . ... . . . . . . .1. . . . . . 93 3. . . . . 86 3. . . . . . . . . . .4. . . . . 65 B Matlab Codes to Plot Fig. . . .. . . . . . 63 Appendices . . . . .. . . . . . . . . . . . . . . . . . 2. 90 3. . . . . . . .. . . . . . 86 3. .. . . .1 Control Design .4 Simulations of Uncertain Single-Pendulum-Type Cranes by ISMC .4. ..5 Second-Order Sliding Mode Control. . . . . . . . . . .. . . .2 Stability Analysis.. . . . . . . . 74 3. . . . . . . . . . . .2 . . . . .xii Contents 2 Crane Mathematic Model . . .. . . .2. . . . . . .1 Control Design . .

. . . . . . .5 Adaptive Control Design Based on Hierarchical Sliding Surfaces. . . . . . . . . .3 Simulation Results . . .1 Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 and 4. . . . . . . . . . . . . . . . . .4. . . .5. . 130 4. . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4. 150 4. . . . . . . . . . . . . . 140 4. . 109 E Simulink Model to Plot Figs.8 . . . . . . . . . .11 and 3. . . . . . . . 4. 100 B Simulink Model to Plot Figs. . . . . . .16.4. . . . . . .3 and 4. 137 4. . . . . . . . . . . .. . . . .11. .. . . . . . . . . . . . . . . . . . . .14 and 4.5. . . . . . . . . 118 4. . 147 4. . . . . . . . . .15 . . . .2 Stability Analysis. . . . . . 98 Appendices . . . . . . . . . .. . . . . . . . . . 117 4. . . . .22 . . . . 127 4. . . . . . . . 136 4. . . . . . 3. . . . . . . . . . . . . . . . . 4. 153 4. . .21 and 4. . 156 A Simulink Model to Plot Figs. . . . 157 C Simulink Model to Plot Figs. . . . .5 and 3. . . . . . . . .6 . . . . . . . . 153 Appendices . . . . . . . . . .6. . . . . . . . . . . . 161 E Simulink Model to Plot Figs. . .12. . . . . . 128 4. . . . . . . . . . .. . . . . . . . . . . . .2 Stability Analysis.1 Control Design . . . . . . . . 119 4. . . . . . . . . . . 107 D Simulink Model to Plot Figs. .12. . .9 . . . . . . . . . . . . . . . . . . . . . . . . 100 A Simulink Model to Plot Figs. . . . . . . . . . . . .4 Simulations of Uncertain Cranes by Second-Order SMC . 4. . . 164 References . . . .5. 3. . . . . . . . . . . . . 4. .17. . . . . . . . . . . . . . 4. . . .1 Control Design . .13. . . . . . . . . . . . . . . . . . . . . . . .1 Problem Description . . .. . . . .4 Combining HSMC .2. . . . .5.7 and 3. . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . .3. . .7. . .3. . . . . . . . . . 4. . .2. . . . . . . . . . . . .3 Simulation Results . . . . . . . .19. . . . . . . . . . . . . . . . . . . . . . . . 115 4 Overhead Crane Control by Hierarchical Sliding Mode . . . . . . . . .2 Stability Analysis. . . . . . . . . . . . . . . .6 HSMC Design for Double-Pendulum-Type Overhead Cranes. . 4. . . . . . . . . . . . . . 132 4. . . . . . 3. . . . . . . . . .3 Simulation Results . . . . . . . . . . 103 C Simulink Model to Plot Figs. . . . 144 4.16 . . . . . . . . . . .6. . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . 4. .15 and 3. . . . . . .1 and 3. . . 150 4. 112 References .2 Stability Analysis. . . 145 4. . . 3. . . . .3. . . . . . . . . . . . . . .6. . . . . . . . 165 . . . . . . . . . . . . . . . . . . .2 Aggregated HSMC. . . . . . . . . .1 Control Design . . . . . . . . . . . . .2. . . . . . . 159 D Simulink Model to Plot Figs. . . . . . . . . . . . . . . 117 4. . . . . . 156 B Simulink Model to Plot Figs. . . . . . . . . . 3. .5. . . . . . .1 Control Design . . . . . .18 and 4. . . . . . . . . .3 Simulation Results . . . . . . . . . . . . . . . . . . .Contents xiii 3.4 . . . . . . . . . . . 4. . . . . . . . . . . . . . 146 4. . . . . . . . . . .6. . . . .3 Simulations of Single-Pendulum-Type Cranes by Second-Order SMC . . . . . . . . . . . . . 121 4. 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. .2 . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . . . . . . 124 4. 96 3. . . . . 135 4. . . .2 Stability Analysis. . . .3 Simulation Results .3 Incremental HSMC . . . . . . .

.4. . . . . . . . . . . . . . 197 6. . . 196 6 Conclusions and Open Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 5. . . . . 198 References . . . . . .xiv Contents 5 Compensator Design Based on Sliding Mode for Uncertain Overhead Cranes . . . . . . .3 Simulation Results . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . 167 5. . . . .2 Compensator Design Based on HSMC . . . . . . . . . . . . . .4 Sliding Mode-Based Neural Compensator Design. . . . . . . . 199 . 168 5. . . . . . . .2 . . . . . . . . . . . . . . . . 170 5. . . . . . . . . . 172 5.3 Simulation Results . . . . . . . . . . . 5. . . . . . . . . . . . . . . .1 Problem Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Stability Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 5. . . . . . . 187 B Simulink Model to Plot Figs.3. . . . . . . . . .4. . . . .2. . . . . .4.2 Extensions and Open Problems . . . . . . . . . . . 197 6. . . .4. . . . . . . 193 References . . . . . . . . . . . . . 5. . . .1 and 5. . .2. . 189 C Simulink Model to Plot Figs. . 5. . . .1 Control Design . . . 187 A Simulink Model to Plot Figs. . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5. . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . 184 Appendices .2 Stability Analysis. . . .3. . . . . . . . . . . . 167 5. . . . . . . . 182 5. . . . . . . . . . . . . . . . . . . . . . .2 Stability Analysis. . . . . . . 180 5. . . . . .3 Simulation Results . . . . . . . .8. . . . . . . . 5. 170 5. . . . . . . .1 Control Design . . . . . .6 . . . . . . . . . . . . .3 Sliding Mode-Based Fuzzy Compensator Design . . . .5 and 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Control Design . . . 173 5. . .1 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 . . . . . . . 179 5. . . . . . . . . . . . . . . . . . . . . .9 and 5. . . . . . . . . . . . . . .

Since cranes are the primary controlled machine in the book. boom cranes. i. and factories throughout the world because of their heavy payload capabilities.e. and steel took over with the coming of the Industrial Revolution [2]. Larger cranes were later developed. According to their primary dynamic properties and the coordinate system that most naturally describes the location of the suspension cable connection point. Yi. warehouses. Cranes are employed to transport heavy loads at shipyards. Then. but cast iron. In the High Middle Ages. and tower cranes [3. iron. employing the use of human tread-wheels. 1. In the late sixth century BC. harbor cranes were introduced to load/unload ships and assist with their construction.Chapter 1 Introduction Abstract This chapter provides necessary background information.. a detailed overview of overhead crane control is presented. cranes can roughly be categorized into three types. The earliest cranes were constructed from wood. permitting the lifting of heavier weights. overhead cranes. Hierarchical Sliding Mode Control for Under-actuated Cranes. Keywords Overhead cranes  Sliding mode control  Crane control 1. Cranes exist in an enormous variety of forms. a brief historical overview of sliding mode control is considered. Qian and J.1 Crane Types A vast number of cranes are used worldwide. © Springer-Verlag Berlin Heidelberg 2016 1 D. construction sites. DOI 10. To review its history and the state-of-the-art research.1007/978-3-662-48417-3_1 .1 were invented by the ancient Greeks and were powered by men or beasts of burden [1]. an introduction to the types of cranes is discussed. the first construction cranes in Fig. These cranes were used for the construction of buildings. The chapter proceeds with some insights into bottleneck issues of control and future research directions. 4].

Boom cranes are often located at building construction sites. The payload is supported by a suspension cable at the end of the boom. illustrated in Fig. a boom rotates around axes both perpendicular and parallel to the ground. 1. mobile overhead cranes are often called gantry cranes.3b. Overhead cranes are com- monly utilized in factories. also known as entitled bridge cranes. warehouses. Different from the other two types of cranes. Sometime. Under this condition. 1.2a. harbors. Such a compact advantage also lends well to being mounted on a mobile base such as trucks. b. ψ is the rotation around the vertical Z axis and θ is the rotation around the horizontal Y axis.2 1 Introduction Fig.1. b. often operate in Cartesian space. 1. and shipyards. overhead cranes travel on a mobile base. .1. In the coordinates. and ships. The schematic repre- sentation of boom cranes is shown in Fig. Boom cranes are typically compact than the other two types of cranes under the condition of similar load carrying capacities. where the trolley moves along a bridge and the motion of the bridge is perpendicular to that of the trolley.3a.1 Ancient greek construction crane 1. tracked vehicles. 1.1 Overhead Cranes The first type is overhead cranes. Boom cranes are usually described by spherical coordinates. This type of cranes. and shipyards. 1.2 Boom Cranes The second major type of cranes is named as boom cranes. Displayed in Fig. the boom cranes can support loads in com- pression.

1.1 Crane Types 3

Fig. 1.2 Overhead cranes.
a An overhead crane in
Jiangsu Province, China.
b Schematic representation of
overhead cranes

1.1.3 Tower Cranes

The third major type of cranes is called tower cranes, like the ones displayed in
Fig. 1.4a, b. For convenience, tower cranes can be described by cylindrical coor-
dinates. From the schematic drawing in Fig. 1.4b, the horizontal jib arm can rotate
around a vertical tower. The payload is supported by a cable from the trolley and
the trolley moves along the jib arm in the radial direction. Tower cranes often give
the best combination of height and lifting capacity. They are commonly found in
the construction of tall buildings because, this type of cranes has the merit of having
a small footprint-to-workspace ratio.
Whatever type the cranes are, a common characteristic among all cranes is that
the payload is supported by a suspension cable. The structural characteristic pro-
vides the basic functionality of cranes, i.e., hoisting and lowering. Meanwhile, it
also presents some challenges. Motion of a crane will definitely result in its payload
oscillation, which is one of the most challenging issues. Inherently, the oscillations
determined by the crane structure are pendulum-type. The oscillations have many

4 1 Introduction

Fig. 1.3 Boom cranes. a A
boom crane in Hubei
Province, China. b Schematic
representation of boom cranes

detrimental effects and a series of drawbacks may be induced, including but not
limited to safety hazards, time delays, and degradation of positioning accuracy.
The ubiquity of cranes versus the propensity to payload oscillations result in
cranes to be the bottleneck in many practical applications. As a result, improving
the control and efficiency of cranes will benefit a lot because of having enormous
economic impact. Significant research attention has been paid to resist payload
oscillations.
Compared to overhead cranes, tower cranes and boom cranes are more chal-
lenging in the aspect of control design because their nonlinear dynamics originate
from the rotational nature of the cranes. This book concentrates on novel structures
of sliding mode control for cranes and adopts crane systems as a research platform.
For the purpose of simplification, only overhead cranes are considered in the rest of
the book.

1.2 Review of Sliding Mode Control 5

Fig. 1.4 Tower cranes. a A
tower crane in Beijing, China.
b Schematic representation of
tower cranes

1.2 Review of Sliding Mode Control

This section reviews fundamental mathematical concepts and common design
methods of the sliding mode control theory. Some numerical examples are illus-
trated in such a manner that readers can understand sliding mode control.

1.2.1 Variable Structure Control

Sliding mode control is also named as sliding mode variable structure control or
variable structure control with sliding mode [5]. From the two aliases, sliding mode

6 1 Introduction

control is often associated with variable structure control and it has a close rela-
tionship with variable structure control. To understand sliding mode control, some
background information about variable structure control is presented.
Variable structure control has a long history. Although the idea evolved from the
pioneering work in Russia in the early 1960s, variable structure control appeared
worldwide only when it was reported by Utikin [5].
Suggested by its name, a variable structure control system has a control law that
deliberately changed during the control process according to some predefined rules.
The rules depend on the state of the system. For the purpose of illustration, consider
a second-order system (1.1).
€xðtÞ ¼ uðtÞ ð1:1Þ

Adopt the feedback control law

uðtÞ ¼ kxðtÞ; ð1:2Þ

where k > 0. The closed-loop system can be demonstrated by means of the
phase-plane method.
Generally speaking, a plot of x_ versus x is an ellipse which depends on the initial
conditions. Figure 1.5 illustrates the system trajectory under the initial conditions
x(0) = 1 and x_ ð0Þ ¼ 0, where k = 0.5 in Fig. 1.5a and k = 5.5 in Fig. 1.5b. In fact,
the control law (1.2) from arbitrary initial conditions cannot force the variables
x_ and x toward the origin. When 0 < k < 1, the ellipse looks like Fig. 1.5a. Shown in
Fig. 1.5b, the ellipse becomes reversed when k > 1. The MATLAB programs of the
example are given in Appendixes A and B.
According to the Lyapunov’s stability scheme, (1.2) can only guarantee that the
second-order system (1.1) is bounded rather than asymptotically stable. Intuitively,
the combined control system may move toward the origin if the two control laws in
Fig. 1.5a, b are pieced together.
For the purpose of illustration, design another alternative control law (1.3) to
piece the two control laws together.

Fig. 1.5 Phase planes. a k = 0.5, b k = 5.5

6 Block diagram of the designed switching control system  k1 xðtÞ if x_x\0 uðtÞ ¼ ð1:3Þ k2 xðtÞ otherwise The block diagram of the designed switching control system is displayed in Fig. The MATLAB programs of the example are given in Appendix C.5 and k2 = 5.3).2 Review of Sliding Mode Control 7 Fig.1. 1.7 An asymptotically stable motion of the system trajectory . the Lyapunov function is defined by VðtÞ ¼ x2 þ x_ 2 ð1:4Þ Fig.7. the intersecting x_ and x-axes divide the phase plane in Cartesian space plane into four quadrants. An asymptotically stable motion is demonstrated in Fig. 1. According to the switching rule in (1.5. To theoretically verify the result. Apply the control law (1. uðtÞ ¼ k1 xðtÞ will be applied to the system in quadrants I and III of the phase plane and uðtÞ ¼ k2 xðtÞ will be implemented on the system in quadrants II and IV of the phase plane. 1. the system trajectory in the phase plane must spiral toward the origin.3) to the system (1. Adopting such a compound control law. where k1 = 0. 1.6.1) from the initial conditions x(0) = 1 and x_ ð0Þ ¼ 0.

x_ Þ\0 In (1. The two control laws result in two different system dynamics (structures) in the phase plane such that the design method is entitled ‘variable structure control.3) originates from intuition and analysis. the control structure is determined at any point (x. x_ Þ is abbreviated as s.2 Sliding Mode Control As mentioned. According to the designed switching function. This fact indicates that the distance from the origin is always decreasing. the switching rule is determined by a function. x_ Þ [ 0 uðtÞ ¼ ð1:6Þ 1 if sðx. (1.2. a stable closed-loop system has been built up by a rule combining two control laws that cannot independently stabilize the system. x_ ) in the phase plane. To systematize the design tool of variable structure control. another variable structure control law can be given as  1 if sðx. V > 0 and V\0.6) can be rewritten as uðtÞ ¼ sgnðsÞ ð1:8Þ .3) into the derivative of V. In a large part. Briefly. the second-order system with the compound control law is of asymptotic stability. As a result. it is just a special case of variable structure control [6].’ 1. sliding mode control has close relevance to variable structure control. x_ Þ ¼ mx þ x_ ð1:7Þ Here. The theoretical deduction coincides with the previous intuitive analysis. Differentiate V with respect to time t and substitute (1. _ In the sense of Lyapunov.1) and (1.4) and (1. The time derivative of V has a form  2x_xð1  k1 Þ if x_x\0 V_ ¼ 2x_x þ 2_x€x ¼ 2_xðx þ uÞ ¼ ð1:5Þ 2x_xð1  k2 Þ if x_x [ 0 From (1. defined as sðx. the switching rule (1. The function is named switching function. A characteristic that variable structure control is different from other control methods is the design of switching rules. m > 0 is scalar and sðx.8 1 Introduction From arbitrary initial conditions. V > 0 exists in (1.4).5).6). Further.

Each cross indicates that the control system is switching between the two different control structures. Figures 1.1) and (1.8). such repeated crosses have a direct effect on the system performance. x_ ¼ mx ð1:14Þ Equation (1.8) into the derivative of V yields V_ ¼ s_s ¼ sðm_x þ €xÞ ¼ sðm_x þ uÞ ¼ s½m_x  sgnðsÞ ð1:12Þ \jsjðmjx_ j  1Þ Equation (1.13). x_ Þ ¼ 0 g ð1:13Þ when m = 1. 1.8 and 1.9 are obtained by a simulink model of MATLAB. Intuitively. 1. which is shown in Appendix D. That is. sgn() denotes a signum function. such switching makes the control input choppy. 1.8) to the second-order system (1. the effect is the high-frequency switching of the control signal and the fluctuation of the system output. Fig. the system trajectory can be constrained to remain on the line L.2 Review of Sliding Mode Control 9 In (1.8.14) means that a straight line crosses the origin in the phase plane.1). In fact. x_ Þ ¼ s2 ð1:11Þ 2 Differentiating V with respect to time t and substituting (1. Under this assumption. .12) indicates the system trajectories on either side of the following line L that will point toward the line when mjx_ j\1. x_ Þ: sðx. defined as 8 < 1 if s [ 0 sgnðsÞ ¼ 0 if s ¼ 0 ð1:9Þ : 1 if s\0 Note that the signum function has a property s sgnðsÞ ¼ jsj ð1:10Þ Apply the control law (1.9. where m is the slope of the line. the Lyapunov function candidate is defined as 1 Vðx.8 illustrates the trajectory slide along the line to the origin. the motion confined to the line L satisfies the differential equation obtained from (1. To consider the stability of this control system in the sense of Lyapunov. Illustrated in Fig. Ls ¼ f ðx. Provided that the switching frequency is infinite.1. the system trajectory repeatedly crosses the line L. Displayed in Fig.

12) from the Lyapunov direct method is referred to as the reachability condition. the trajectory moving toward the surface is called the reaching mode. b control signal u As a result. The line L is entitled as the sliding surface.9 System performance when m = 1. the trajectory sliding along the surface is termed the ideal sliding mode.10 1 Introduction Fig. which are usually written as s_s\0 ð1:15Þ or equivalently lim s_ \0 and lim s_ [ 0 ð1:16Þ s!0þ s!0 . 1. 1. the high-frequency switching of the control signal is described by chattering and the condition in (1. a System output x.8 Phase portrait of a sliding motion of the system when m = 1 Fig. some terms to describe sliding mode control can be drawn.

According to the equivalent control law.1) and (1. Then. the equivalent control is marked by ueq. (1. Consider the system (1.. Note that the equivalent control action is not the real control signal applied to the system but may be treated as the control signal applied on average. (1. Usually. To illustrate the superior and design of the equivalent-control-based sliding mode control. both of the controls in (1.7).1).20). These linguistic descriptions can be formulated as s ¼ 0 and s_ ¼ 0 ð1:17Þ From (1.15).2.7).7). uðtÞ ¼ ueq þ usw ð1:20Þ In such a design. Substituting (1. and adopt the sliding mode control law (1.1). the existence of the sliding mode is guaranteed.19) mains the sliding motion of the second-order system on the line L.15) yields s_s ¼ sðm_x þ €xÞ ð1:21Þ ¼ sðm_x þ uðtÞÞ ¼ sðm_x þ ueq þ usw Þ Consider (1. equivalent control and switching control. the original control signal can be divided into two parts. i.21) yields s_s ¼ susw ð1:22Þ . Suppose that the system trajectory reaches the sliding surface at time ts and that the sliding mode exists thereafter. design the sliding surface (1. design the sliding surface (1.20) into the left of (1.19) is named equivalent control.19).3 Equivalent-Control-Based Sliding Mode Control Consider the system (1. Then.2 Review of Sliding Mode Control 11 1.1). (1.18) can be obtained s_ ¼ m_x þ uðtÞ ð1:18Þ Since s_ ¼ 0 for all t > ts.20) cooperate with each other to generate the real control signal in the reaching mode stage but the visual equivalent control is solely applied to the system in the sliding mode stage. the fol- lowing case study is conducted.8).1. rearranging (1. The switching control law usw can be deduced from the reachability condition (1. and (1.19).7) and adopt the sliding mode control law (1. the control law (1. uðtÞ ¼ m_x ðt  ts Þ ð1:19Þ The control law (1.e.

8) is to ensure the phase tra- jectory moves toward and is forced to remain on the sliding surface.12 1 Introduction From the reachability condition (1. As mentioned. (1. the system perfor- mance under the equivalent-control-based sliding mode control law is displayed in Fig. Further.10 and 1. 1. the phase trajectory in Fig. Then.8). the equivalent-control-based sliding mode control law (1. which is shown in Appendix E. ð1:23Þ where k > 0 are scalar. some simulation results are shown in Figs.20) of the second-order system (1.9. the value of k in (1. Such an equivalent-control-based design can dramatically reduce the chattering. The purpose Fig.11 are obtained by a simulink model of MATLAB.8). 1. the curves are smoother and have less fluctuation. 1. Compared with the results in Fig.15). Compared with Figs.11. the equivalent-control-based sliding mode control law (1.22) becomes s_s\0 ð1:24Þ Finally.7). Figures 1.8 and 1.1) has the form uðtÞ ¼ m_x  k sgnðsÞ ð1:25Þ Compared with the sliding mode control law (1.25) is kept unchanged from k = 1 and the parameter of the sliding surface is also designed by m = 1 in (1. To demonstrate the feasibility of the equivalent-control-based sliding mode control. 1. 1. the purpose of the control law (1. For the purpose of comparison with the precious control law (1. define usw ¼ k sgnðsÞ.10 is much smooth and the system performance is apparently improved. where the visual equivalent control law is continuous. 1.25) consists of two parts.10 and 1.10.10 Phase portrait of a sliding motion of the system when m = 1 and k = 1 .11.

2 Review of Sliding Mode Control 13 Fig.1. a System output x. . has the form s h¼ .11 System performance when m = 1 and k = 1. The idea of equivalent-control-based sliding mode control divides the whole control action into two parts. ð1:26Þ jsj þ d where s denotes the sliding surface variable and δ is a small positive constant. many methods are implemented. Note that the chattering is an inherent drawback of sliding mode control. due to physical limitations in real world systems. b control signal u results in a dramatic switching of the control signal generated by (1. where the smoothing function method is representative. The chattering problem is con- sidered as a major and inherent drawback for sliding mode control. which can reduce the chattering phenomenon and improve the system performance. To reduce chattering. the oscillation is called chattering. However. As mentioned. where the control action on the sliding surface is continuous. The smoothing function is a means for eliminating chattering while trying to preserve good robustness properties around the sliding surface. The equivalent- control-based sliding mode control design fills the gap and such a control design makes the sliding mode control theory applicable. the methodology of sliding mode control requires an infinitely fast switching mechanism. 1. which is also a gap between sliding mode control and real applications. the chattering may cause unpredictable instability. There is a large class of smoothing functions available to convert the discontinuous control laws.4 Chattering Reduction Theoretically. which is used for single input single output systems. 1.8). directly applying the above control design will always lead to oscillations of the control signal in some vicinity of the sliding surface.2. For a gain switching system. One of them.

Compare Fig.14 1 Introduction Fig. the control signal becomes smooth in the dynamic process by introducing the smoothing function in Fig. 1. The MATLAB programs of the example are given in Appendix F.11 and 1. the system is effectively a high gain feedback system and possesses the properties of certain classes of disturbance rejection and parameter invariance. But the control signals in Figs. From Fig.12.25). 1. but it decreases the robustness of sliding mode control.12 A smoothing function of the signum function Numerical solutions of (1. Adopt the control (1.13 with Fig. 1.12.1.14. 1. On the other hand.27) for the system (1. 1.10. smoother the curve is.27).14. The two curves of phase trajectories almost make no difference in the phase plane. 1. bigger the value of δ is. degenerates sliding mode into quasi-sliding mode. the new control law can be rewritten as s uðtÞ ¼ m_x  k ð1:27Þ j sj þ d In (1. In Fig. Thus. The replacement can improve the performance of control signal. the control signal severely jumps back and forth at about time t = 1 s because the signum function is triggered by the sliding motion in the phase plane.26) replaces sgn(s) in (1. .14 illustrate the effects of the smoothing function method.13 and 1. 1. 1. Concerning the alternative.11. 1. and loses guaranteed system stability. which are obtained by a simulink model of MATLAB in Appendix G. The system phase trajectory and the control signal are illustrated in Figs.25).1). (1. such an alternative will only effect the second term in the control law given by Eq. where the values of m and k are kept unchanged from m = 1 and k = 1 and the value of δ is 0. in the region around the sliding surface determined by the value of δ. (1. The variable δ can be used to trade-off the requirement of maintaining ideal performance with that of ensuring a smooth control action. As far as the equivalent-control-based sliding mode control is concerned.26) are shown in Fig. It can be visualized that the function θ tends point-wise to the signum function as δ → 0.

accuracy. stability. 1. . the performance of the control signal can be improved by increasing the value of δ. a System output x. But a series of drawbacks may be induced by the alternative smoothing function method. k = 1 and δ = 0.14 System performance when m = 1. A more general frame will need to be established.2 Review of Sliding Mode Control 15 Fig. The state space approach provides a possible choice for accomplishing this task. 1. phase plane analysis is hardly utilized to analyze multivariable systems and high-order systems. b control signal u In Fig.1.14. such as performance. 1.1 Fig.1. The value of δ must be determined by a trade-off among the performance indexes.5 Sliding Mode Control Design for State Space Model Although it is an intuitive and effective tool.2. and so on. 1.13 Phase portrait of a sliding motion of the system when when m = 1. k = 1 and δ = 0.

The states of the system can be represented as a vector within that space. 3–5. x_ ¼ Ax þ Bu is named state equations. Define a vector as x . i.1) can be rearranged in the form of state space. D 2 <qp is termed feed-forward matrix. B 2 <np is called input matrix. u 2 <p is termed control vector. take the second-order system (1. (1. Further. sliding mode control. the general expression of (1.29) is written as x_ ¼ Ax þ Bu ð1:32Þ . can be described in the form of matrix as well. the more general state space represen- tation of a linear continuous time-invariant system with p inputs.30). Concerning (1.16 1 Introduction A state space representation is a mathematical model of a physical system. These state variables are related by a set of first-order differential equations. From the viewpoint of control theory.28) depicts the state space model of a continuous time-invariant system. which is a set of input.. the sliding surface can be expressed in matrix form as s ¼ cT x ð1:30Þ here c ¼ ½ m 1 T . q outputs. C 2 <qn is called output matrix. output and state variables.e. Thus.1) into consid- erations. and n state variables can be written in the following form x_ ¼ Ax þ Bu ð1:31Þ y ¼ Cx þ Du Here x 2 <n is called state vector. Since this book considers control design by the state-feedback-based sliding mode control. y 2 <q is named output vector. The so-called state space refers to the space whose axes consist of all or part of the state variables. equivalent-control-based sliding mode control and chattering reduction. the aforementioned design methods. only the state equations of the general state space representation are employed in the rest chapters. and y ¼ Cx þ Du is called output equations. whose expression is as follows:     0 1 0 x_ ¼ xþ u ð1:29Þ 0 0 1 Further. ½ x x_ T ð1:28Þ Then.28) and (1. the system (1. A 2 <nn is termed input matrix. For the purpose of illustration. The details are exhibited in Chaps.

One type is matched and the other type is unmatched. where matched uncertainty enters into a control system by the control channel.1.36) into (1. The phase portrait in Fig. Robust methods aim to achieve robust performance and stability in the presence of bounded modeling errors. dðx. which generates a uniformly distributed pseudorandom number in the open interval ð 0 1 Þ. 1. To demonstrate the effects of matched uncertainty on the system performance. Although it is not categorized into traditional robust control methods. define   0 dðx. sliding mode control is also strongly robust. model uncertainties and random external disturbances [7]. Sliding mode control exhibits its robustness by invariance. complementing an uncertain term in (1. (1. tÞ ð1:35Þ 0 0 1 In (1. As far as it is concerned.32) into the derivative of s yield s_ ¼ cT x_ ¼ cT ðAx þ BuÞ ð1:33Þ The general expression of the equivalent control law can be written as  1 ueq ¼  cT B cT Ax ð1:34Þ 1.35).2.35) yields .29) are called nominal systems. robust control is developed as a branch of control theory.10 shows the sliding motion remains on the sliding surface as if the sliding surface were the system dynamics. Substituting (1. Uncertainty includes (but is not limited to) parameter fluctuations. tÞ is the distance term and the original Eq.6 Robustness Against Uncertainties A real process is evidently and inevitably uncertain. uncertainty covers two types.29) yields     0 1 0 x_ ¼ xþ u þ dðx. The characteristic is called invariance.30) with respect to time t and substituting (1. To deal with uncertainty. tÞ ¼  0:1  randðÞ. For instance. Concerning sliding mode control. ð1:36Þ 1 where randðÞ is a MATLAB command [8]. which is also the most attractive feature of sliding mode control.2 Review of Sliding Mode Control 17 Differentiating (1. its invariance property is robust enough to suppress the type of matched uncertainty and it can guarantee the sliding mode is maintained as if there was no matched uncertainty.

it undoubtedly belongs to matched uncertainty.15) yield s_s ¼ scT x_ ð1:38Þ Substituting (1.37) into (1. Consequently.18 1 Introduction     0 1 0 x_ ¼ xþ ½u þ 0:1  randðÞ ð1:37Þ 0 0 1 From (1.39) by the general expression (1.38) yields      0 1 0 s_s ¼ sc T xþ ½u þ randðÞ ð1:39Þ 0 0 1 Consider the equivalent-control-based sliding mode control law.30). we resort to the reachability condition of sliding mode (1. To verify the system stability. Substituting the equivalent-control-based sliding mode control law into (1.1) by the control channel.37). differentiating (1.15).30) with respect to time t and substi- tuting the derivative of s into (1. Designing the sliding surface (1.39) and rearranging (1. the disturbance term enters the system (1.32) yield    s_s ¼ scT Ax þ B ueq þ usw þ B  0:1  randðÞ n  1 o ¼ scT Ax þ B  cT B cT Ax þ usw þ B  0:1  randðÞ ð1:40Þ .

k is a design parameter. which are obtained by a simulink model of MATLAB in Appendix H.10. 1.15 and 1. Some numerical simulation results from the initial conditions x(0) = 1 and x_ ð0Þ ¼ 0 are displayed in Figs. Compared with Fig. The trajectory frequently crosses the sliding surface because of the existence of matched uncertainty. the reachability condition is satisfied if and only if k [ supk0:1  randðÞk ð1:41Þ supk0:1  randðÞk ¼ 0:1 in (1. 1. 1. This property is termed invariance of sliding mode. Once the trajectory deviates from the sliding surface. the switching control law and the equivalent control law will draw the trajectory toward the surface together. The switching in Fig. adopt the equivalent-control-based sliding mode control design (1.16a. the system output can eventually tend to be .16.15 has no ideal sliding mode. ¼ s cT Busw þ cT B  0:1  randðÞ ¼ scT B½usw þ 0:1  randðÞ Let usw ¼ k sgnðsÞ. For the purpose of illustration. 1. which is only related to the boundary of uncertainty rather than any physical parameters.20) and keep the parameters of the sliding surface and the switching control law unchanged from the precious example.41).16b reflects this process. indicating we can choose any k [ 0:1 as the parameter of the switching control law. the phase portrait in the phase plane in Fig. Illustrated in Fig. 1.

38) into (1. The equivalent-control- based sliding mode control design (1. 1. to resist the uncertainty.20) is still adopted. the uncertain term is unmatched.11b. define   1 dðx.1) by the control channel. a System output x.16c. 1.43).32) yields       0 1 0 1 x_ ¼ xþ uþ  0:1  randðÞ ð1:43Þ 0 0 1 0 From (1. To demonstrate the effects of unmatched uncertainty. c control signal u asymptotically stable.16 Performance of the system with matched uncertainty when m = 1 and k = 1. the disturbance term can hardly enter the system (1. 1. b sliding surface s.1. 1.15 Phase portrait of a sliding motion of the system with matched uncertainty when m = 1 and k = 1 Fig.2 Review of Sliding Mode Control 19 Fig. the chattering phenomenon of the control signal is severe in Fig. tÞ ¼  0:1  randðÞ ð1:42Þ 0 Substituting (1. Since the unmatched . Compared with Fig. Consequently.

15) yield s_s ¼ scT x_ ð1:44Þ Substituting (1.30) with respect to time t and substituting the derivative of s into reachability condition (1.32) yield       1 s_s ¼ sc Ax þ B ueq þ usw þ T  0:1  randðÞ 0   1  1 ¼ scT Ax þ B  cT B cT Ax þ usw þ  0:1  randðÞ ð1:46Þ 0     1 ¼ s cT Busw þ cT  0:1  randðÞ 0 Let usw ¼ k sgnðsÞ. Substituting the equivalent-control-based sliding mode control law into (1. The system output curve in Fig. which is illustrated in Figs.18. Designing the sliding surface (1. the reachability condition is satisfied if and only if     1  kBk k [   0  0:1  randðÞ   ð1:47Þ Equation (1. . To illustrate the effects of unmatched certainty. Compared with Figs.47) indicates that the equivalent-control-based sliding mode control system can be of asymptotic stability in the presence of unmatched uncertainty under the assumption that the unmatched uncertainty has a known boundary. the system stability should be evaluated at first. 1.30).17 illustrates the system trajectory in the phase place.43) into (1. some numerical simulation results are in Figs. On the other hand. 1.18a also supports this point.20 1 Introduction uncertainty cannot be suppressed by the invariance property. the switching control law are kept unchanged from the precious example and the initial conditions x(0) = 1 and x_ ð0Þ ¼ 0. the system with unmatched uncertainty cannot tend to be stable until 40 s. where the parameters of the sliding surface.45) by the general expression (1.20. differentiating (1. In fact.17 and 1. Although it tends to the origin. 1.45) and rearranging (1.19 and 1.15 and 1.15. 1.44) yields        0 1 0 1 s_s ¼ sc T xþ uþ  0:1  randðÞ ð1:45Þ 0 0 1 0 Consider the equivalent-control-based sliding mode control method. the system with matched uncertainty has been stable via the same control law and the same controller parameters in Fig. the trajectory is still far away from the desired origin till 20 s. 1.

b sliding surface s.19 Phase portrait of a sliding motion of the system with unmatched uncertainty at t = 40 s .2 Review of Sliding Mode Control 21 Fig.17 Phase portrait of a sliding motion of the system with unmatched uncertainty when m = 1 and k = 1 Fig. c control signal u Fig. 1. 1. a System output x.18 Performance of the system with unmatched uncertainty when m = 1 and k = 1.1. 1.

i. the unmatched certainty affects the system stability by coupling the physical parameters. The order of a sliding mode [6] represents the smoothness degree associated to the motion constrained on the sliding manifold.7 Sliding Order and Sliding Sets A function is of class Ck. Under either kind of uncertainty. t) = 0 is referred as (k + 1)th order sliding mode. defined as follows. The sliding mode control law tends to make the system trajectory move toward the sliding surface by the sliding mode control law. f is continuous with its derivatives up to the order k. Definition 1. On the other hand. Note that the sliding mode of order one corresponds to a sliding variable of class C0. . the sliding mode control system is of asymptotically stability under both matched and unmatched uncertainties. 1. if the sliding surface variable s is of class Ck.19 and 1. 1. c control signal u From Figs. the sliding mode will deviate from the sliding surface.20 Control performance of the system with unmatched uncertainty at t = 40 s. 1. Consequently. both kinds of uncertainties affect the sliding modes and deteriorate the chattering phenomena. t) whose vanishing defines the equations of the sliding mode.2.22 1 Introduction Fig.1 The sliding order r is the number of continuous total derivatives (including the zero one) of the function s = s(x. Correspondingly. with discontinuous derivative. 1.18.. 1. The sliding mode control system can realize the control task quickly against the matched certainty because the matched certainty has no relationship with the physical parameters and can directly affect the system stability.20.17.e. then the sliding mode on the surface s(x. which delays the system response. a System output x. when x denotes the system states. b sliding surface s.

with a diversity of published control techniques.e.1 Open-Loop Control 1. These methods can avoid the sensing problems of closed-loop control methods.2. The similar control problems were also investigated by the Pontryagin’s maximum principle [10] and some evolutionary algorithms [11]. sðÞ ¼ s_ ðÞ ¼    ¼ sðr1Þ ðÞ ¼ 0 ð1:48Þ From Definitions 1. i. Ross and Fahroo [15] proved that the optimal motion planning problem was equivalent to a classic Bolza problem of the calculus of variations and their results were exhibited by the 2-D crane model. also named motion planning.3.1. Since the book focuses the control problems of cranes. as well as dynamic crane models. which does not consider crane models. 1. The control problem is typically formulated as a minimum time optimization problem.3. where the open-loop methods can be further divided into optimal trajectory planning and command shaping.2 The sliding set of order r associated to the manifold s(x. Lee and his research division solved several motion planning-based methods to realize the high-performance anti-swing control of overhead cranes. t) = 0 is defined by the equalities. In Chap.2 Review of Sliding Mode Control 23 Definition 1. 1.3 A Review of Crane Control Crane control is a mature research area.13) is just the first-order sliding mode. .. the sliding set in (1. 3. seek to elimi- nate vibration by using preplanned trajectories to move the crane through the workspace. crane control methods can be divided into open-and closed-loop methods. and fusing the motion planning method on the sliding mode control [14]. a review of crane control is presented in this section. As most control problems. The first to prose an optimal control strategy for automatic crane operation was Field [9]. solving the motion planning problem through a kinematic problem [12]. designing the motion planning method by means of the model-following anti-swing control law [13].1 and 1. some typical sliding mode control methods will be introduced one by one though control problems of overhead cranes.1 Optimal Trajectory Planning Optimal trajectory planning methods. but also releases some pressure on control for the sequential strategy.1. are reported in the last two decades. A lot of efforts in crane control. subject to vibration constraints. Optimal design of the transportation trajectory not only improves the control performance in terms of efficiency or safety.

The amplitude and time locations of the impulses are obtained from the system’s natural frequencies and damping ratios. dynamic programming [18] and linear programming [19]. To employ the input shaping method. The input shaping technology is widely used. Programming is an effective optimization tool. So far. To deal with the issue. the method has been implemented on several large bridge cranes at nuclear facilities [23].e. Unlike the optimal trajectory planning. The primary restriction of the optimal trajectory planning method is that the desired motion of cranes must be known in advance. the incremental coordination method is employed to plan the motion of two or more cranes. an input shaper should be designed. AlBahnassi and Hammad [20] developed a frame- work for near real-time motion planning of cranes. That is.3. i. in addition to the initial conditions of the maneuver. where the trajectory was tuned by an iterative learning strategy to guarantee accurate trolley positioning. Generally speaking. Input shaping not only achieves crane control directly. Blajer and Kolodziejczyk [16] first sketched a trajectory by a series of points in the work space and then approximated the trajectory by spline functions. Crane control research in the input shaping field involves developing shapers to minimize the sway of crane payloads. where the kinematics and the geometrical constraints of cranes are taken into consideration.1. where some rigorous geometric analysis were utilized to address the coupling behavior between the actuated trolley motion and the underactuated payload swing.2 Input Shaping Input shaping is another open-loop control technology for cranes. The framework can assist crane operators to replan safe paths in near real time. which makes the method non robust. vibration caused by a part of the impulses is canceled by vibration caused by other parts of the impulses. they [22] considered an offline trolley trajectory planning method for underactuated overhead cranes. The use of input shaping dramatically reduces the time to the completion of move operations and greatly increases the . a 10-ton crane at Georgia Tech [24]. Another major drawback is that generating the optimal profiles can become computationally expensive because of the kinematic coupling behavior between the underactuated payload’s swing and the actuated trolley’s motion. 1. To coordinate multiple construction cranes [17]. Further. which is composed of a series of impulses. tower cranes [25] as well as coordinate measurement machines [26]. which can dra- matically reduce motion-induced payload oscillations by intelligently shaping the reference commands.24 1 Introduction To deal with the crane control problems in cluttered work environment. Several programming methods have been applied to plan crane trajectories. but also it can improve crane operator performance [27]. trajectory planning is offline. input shaping is an online method and it can be applied in real time. Fang and his colleagues [21] investigated a kinematic coupling-based offline trajectory planning method for overhead cranes. These impulses can cancel vibration..

convolved and simultaneous two-mode input shapers [31]. However.. backstepping. which could be treated as the rudiment of the current input shaping method. and negative-impulse-included shaper [33]. Publications concerning crane control by input shaping have resulted in dozens of papers. Yamada et al.3 A Review of Crane Control 25 safety of the crane by allowing for a smaller danger zone around the crane. Concerning crane control. Lin [30] designed an input shaping method by a pole/zero placement approach. energy. For the purpose of illustration. Singhose and his colleagues also suggested several shaping methods. Especially. [28] first proposed a shaping-type control strategy to control cranes in 1965.e. Direct applications of many traditional nonlinear control methods also suffer from such structure. Alsop et al. some novel shaping methods are often reported. expert-system-based shaper [32]. For the purpose of illustration. where the control command was generated by Pontryagin’s maximum principle to achieve minimum transfer time. nonlinear input shaping [34]. meaning that they are hard to resist any error without special design. shaping design considering limited-state constraints [36].. Thereafter. trolley position. 1. sliding mode. [29] proposed an optimal input shaping method. . the survey of closed-loop control for cranes is demonstrated in Fig.1. many control methods based on feedback linearization.. the draw- back becomes deteriorative for some types of double-pendulum cranes because there are severe residual vibrations even for a small modeling error. which uses information about the current states of the system (e.3. the control input is the force or torque applied to the trolley in order to suppress oscillations due to the acceleration and deceleration of the trolley. The complexity of the crane control field originates from the crane structure that the number of actuators is less than the degree-of-freedom to be controlled. payload swing angle. input shapers are located at the feed-forward channel. Recently. etc. the survey of open-loop control for cranes is demonstrated in Fig.22. a large research effort has been directed at crane control applications via the input shaping method owing to the effectiveness of input shaping.21. where the pole/zero placement method could improve system perfor- mance and simplify design procedures. and fuzzy logic have been developed for crane control. frequency-modulation input shaping [37].2 Closed-Loop Control Closed-loop crane control aims to find a feedback control law that stabilizes the system in the presence of various uncertainties and external disturbances. i. 1. 1. Recently.) to generate commands that drive the system toward the desired state.g.e. This section provides a view to highlight some of these popular control methods. i. graphical-approach-based input shaping design [35].

linear quadratic regulator (LQR) control by Grassin et al. [39]. state feedback control by Nguyen [41]. com- bination of input shaping and linear control by Yoon et al. i. In particular. Based on the linearized crane model.2. Hazlerigg [38] was one of the first to propose this method in 1972.26 1 Introduction Fig. 1. linear control methods are sensitive to changes in the cable length because they are invariably tuned to counter the effects of the natural .e. and parameter optimization of linear controller by Hamalainen et al.1 Linear Control Linear control calls for linearized crane model.21 Survey diagram of open-loop control for cranes 1. [43]. However. Since then. [42]. a variety of linear control methods has been applied to crane control practice. the crane model is inherently nonlinear and the linearized method cannot provide the sufficient accuracy of information about position error and load swing. These methods include some classic control design.. proportional-integral and proportional −derivative controllers by Lee [40].3.

[45] investigated the feasibility of partial feedback linearization for crane control problems. respectively.2. [49] designed a novel control method on the basis of partial feedback linearization. and the friction of track will also reduce the performance of linearized crane control systems. Sun and Fang [48] reported their results about the applications of partial feedback linearization with saturation constraint. Olfati-Saber [44] introduced a global change of coordinates that can decouple these two subsystems but leave the linear subsystem invariant. Zhang et al. where the overall crane system is divided into two subsystems. and the control method can be easily applied to the new form of the system. There are two partial feedback linearization techniques presented as collocated and non-collocated partial feedback linearization methods. Wu et al. the hit.2 Partial Feedback Linearization A useful technique for crane control is called partial feedback linearization. The method has been extensively applied to several types of cranes. 47] extended the method to overhead cranes with varying cable length and 3-Doverhead cranes. the uncertain nonlinear factors such as the wind.1.3 A Review of Crane Control 27 Fig. Besides. which is a method providing a natural global change of coordinates that transforms the system into a strict feedback form.3.22 Survey diagram of closed-loop control for cranes frequency of the cable-payload at a single cable length. 1. 1. . Tuan [46.

convey cranes by Collado et al. Especially. Roughly. 60] for designing stabilizing controls for a special class of nonlinear dynamical systems.e. 1. the backstepping technique often combines other . Since then. One limitation of passivity-based control is that it is only applicable to the systems with relative degree less than two [58].28 1 Introduction Summarily.3 Energy-Based Methods In light of the Lyapunov’s stability theory. i.e. These systems are built from subsystems that radiate out from an irreducible subsystem that can be stabilized using some other method. Thereafter. such as storage energy or dissipation plays a fundamental role in the stability analysis and performance. d’ Andréa-Novel and Coron [61] analyzed the exponential stabilization of an overhead crane with flexible cable via a back- stepping approach. Alli and Singh [53] developed an optimal passive control approach for flexible structures like cranes. Stability prop- erties based on the Lyapunov theory can be easily studied for the obtained closed-loop systems. cranes with heavy chains and payload by Thull et al. Pointed out by Maschke et al. Because of this recursive structure. passivity-based techniques have been used in crane control problems. they are always utilized as an initial simplifying step for crane control problems. a designer can start the design process at the known-stable system and “back out” new controllers that progressively stabilize each outer subsystem. Proven by Karkoub and Zribi [51]. The process terminates when the final external control is reached. cranes are inherently one of the isolated standard Hamiltonian systems.. [56]. i. [55].3. double-pendulum overhead cranes by Guo et al. there are two kinds of energy-based control methods for cranes. respectively. To overcome this limitation. Fang and his colleagues [57] proved the equivalence relation between passivity-based control and a proportional–derivative control concerning the special case of crane control problems. the advantage of collocated and non-collocated partial feedback linearization methods is a conceptual and structural simplification of control problem. They are presented as passivity-based control and backstepping-based control. it is well-known that the total energy is a suitable Lyapunov function to study the sta- bility of the trivial equilibrium of an isolated standard Hamiltonian system. Sun and Fang [52] presented some primary results on cranes by energy-based analysis. [54]. Concerning crane control problems.2. this process is entitled backstepping. A nice feature of the passivity-based control design is the physical meaning of the resulting control laws and the concepts. Backstepping is developed circa 1990 by Kokotovic [59.. a technique called backstepping is proposed to transform the system into a new recursive form where passivity-based control can be easily applied to. Hence. a Lyapunov function candidate is the foundation to analyze the system stability. [50].

Robust control is a branch of control theory. In the past few decades. [67] investigated the model reference adaptive control method for overhead cranes.. Besides. the procedure of backstepping becomes very complicated and implementation of such a control design in practical applications may be unrealistic. and external dis- turbance. according to the traditional adaptive control approaches. Adaptive control [65] is the control method used by a controller which must adapt to a controlled system with parameters which vary. adaptive backstepping control by Cao et al. Hurteau and Desantis [66] suggested a simplified adaptive control law for cranes. which aims to achieve robust performance and stability in the presence of bounded modeling errors. external distance [72]. i.e. parameter uncertainty. when the degree-of-freedom of cranes increases. Ackermann [80] investigated the parameter space . 64]. [62]. However. to name but a few. adaptive sliding mode fuzzy control [76].3 A Review of Crane Control 29 control methods to realize crane control problems. i. adaptive sliding mode control [74]. which does not need a priori information about the bounds on these uncertain or time-varying parameters. and unknown parameters [73] are also reported. d’Andrea-Novel and Boustany [68.2. where all the control system parameters were not needed for the controller design in a priori. adaptive coupling control [77]. there are typical discrepancies between a practical system and its theoretical model because of unmodelled dynamics. respectively. for instance. Butler et al. 1. [63. Such a technique is developed under the assumption that uncertain parameters or disturbances are found within some compact set.1. An adaptive control system utilizes online identification of either system parameter or controller parameter. For any control design. backstepping sliding mode control by Tsai et al.e.3.4 Robust Control and Adaptive Control Uncertainty is a common yet intractable issue for crane control. The theory of robust control began in the late 1970s and early 1980s and soon developed a number of techniques for dealing with bounded system uncertainty [79]. 69] reported two indirect adaptive control methods for cranes. some adaptive design approaches against cable length changes [71]. a lot of effort has been directed in adaptability and robustness to these factors. adaptive control and robust control are developed to solve the adaptability and robustness issues of uncertainty compensation. there are still some research results by combining adaptive control and other control methods. or are initially uncertain. Recently. motion planning-based adaptive control [75]. where their adaptive controller was made of a state regulator block and a gain tuning module. These adaptive control approaches consider their control design in the sense of Lyapunov. As two main branches.. double- pendulum-type overhead cranes. and adaptive control based on fuzzy cerebellar model articulation controller [78]. Yang and Yang [70] developed an adaptive control method for 3-D cranes.

Owing to the nature of these two control structures. operator-based robust right coprime factorization design methods by Wen et al. [86–88]. the design of an adaptive control system [65] is concerned with the adaptive control law changing itself to suppress all uncertainties.. [81] developed a multimode vibration control method to suppress crane vibrations. Although both the control approaches aim to resist the uncertainties of crane systems. The idea of robust control [79] can guarantee a designed robust control system to be insensitive to all uncertainties by using a fixed control structure. Kar et al.5 Predictive Control Model predictive control is an advanced control method that has been in use in the process industries. indicating that the robust control law need not be changed if the changes are within given bounds. developed by Duncan McFarlane and Keith Glover of Cambridge University. This kind of design can guarantee that the system will not greatly deviate from expected trajectories when disturbances enter the system. [92] first proposed a model predictive control method for cranes. Since model predictive controllers rely on dynamic models of the process. i. since the 1980s. On the other hand. linear empirical models obtained by system identification but finding a closed-loop control law for cranes is usually quite complex because of their highly nonlinear equations of motion. [94] investigated the application of model predictive control of boom cranes. where an optimizer was employed to find an open-loop solution at each sampling interval for a given horizon. 1. wave-based robust control by Yang and O’Connor [82].e. robust control is only suitable for dealing with small uncertainty. As a result. linear matrix inequality-based optimal robust control [89]. most often. Besides the classic H∞ loop-shaping robust control. [83]. there are noreports on the aspect of . Adopting H∞ loop-shaping. and model predictive controller was located at feedback channel to compensate for the optimizer.2. Arnold et al. and gain scheduling-based robust control [90]. the motivations behind the two approaches are different from each other. in chemical plants and oil refineries. Another important example of a robust control technique is H∞ loop-shaping. to name but a few. hybrid control on the basis of proportional- differential control and H∞ scheme [84]. it has also been used in crane control problems [91]. but it is sen- sitive to unstructured uncertainty. Kimiaghalam et al.3. partial-state-feedback-based robust control by Uchiyama [85]. Deng and Becerra [93] explored the constrained predictive control for cranes. pole-placement-based robust control by Uchiyama et al. The technique [79]. such as cable length and payload mass. minimizes the sensitivity of a system over its frequency spectrum. while adaptive control is suitable for a wide range of parameter variation. In recent years.30 1 Introduction design of robust control and illustrated the use of the design tool by a crane control system. other robust control methods also are developed against the changes of cable length and payload mass. Subsequently.

the model predictive control algorithm has the ability to anticipate future events and it can take control actions accordingly.. so that their experience can be used in the control design. machine and computer control are accurate and efficient but unadapt- able. model predictive control of the crane becomes active again. As far as fuzzy logic is concerned. evolutionary computation. fuzzy logic. genetic algorithms. imprecise. fuzzy crane control problems were first touched by some Japanese scholars [106. main contributions exhibit how to apply the theory of fuzzy logic to the crane control problems. and genetic algorithms are often employed to optimize controller parameters and con- troller structures and they are rare to directly generate control commands. and so on [103]. Suggested by its name. solving constraint problems of crane control [96.2. fuzzy logic among these intelligent techniques has also been paid more attention [104]. For the duration of these pioneering researches. but also it can directly generate control commands applying to control systems.e. Some novel approaches on different sides of model predictive control for cranes are reported. while keeping future timeslots in account. 97]..3. In the end of 1980s. Fuzzy logic has the advantage [105] that the solution to a problem can be cast in terms that human operators can understand. fuzzy logic not only has the ability to optimize control systems. but rather as “partially true”. neural networks.1. Bayesian probability. i. The related researches were continued till 1990s [108–111]. and fuzzy. there exists a mismatch between human and machine control because human factors result in uncertain.e. Fuzzy logic is widely used in machine control. main challenges of model predictive control are that the control algorithm highly depends on system model and that the model predictive control algorithm has no guaranteed stability. Zadeh of the University of California at Berkeley in 1965. In contrast. This is achieved by opti- mizing a finite time-horizon. and nonlinear model predictive control against crane nonlinearities [99–102]. 1. .6 Intelligent Control Intelligent control is a class of control techniques that use a variety of intelligent computing approaches. The advantage of model predictive control [91] is the fact that it allows the current timeslot to be optimized. i. Fuzzy logic is a bridge to make machines intelligent enabling them to behave in a fuzzy manner like humans [58]. Consequently. The term “fuzzy” refers to the fact that the logic involved can deal with concepts that cannot be expressed as “true” or “false”. Meanwhile. Thereafter. such as neural networks. [95] presented a time-optimal formulation within the model predictive control framework. However.3 A Review of Crane Control 31 model predictive control of cranes until Van den Broeck et al. 107]. designing state observers to approximate unmeasurable states [98]. to name but a few. Fuzzy logic [105] was first proposed by Lotfi A. but only implementing the current timeslot. Among various approaches. evolutionary computation. This makes it easier to mechanize tasks that are already successfully per- formed by humans.

156–160]. Thereafter. and control practice of fuzzy crane systems [113]. . One type is to inves- tigate various siding mode design methods for cranes. The other type is to fuse other control methods to improve performance of sliding mode-based control systems for cranes. Lyapunov redesign [133]. analysis of the crane dynamics [109. and adaptive sliding mode control by some researchers [74. high-order sliding mode control by Chen and Saif [142]. and iterative and evolutionary optimization [134] can be found.3. optimization of fuzzy systems [112]. Further. where fuzzy logic is employed to approximate the nonlinear crane model and the final control command is generated by other control methods. incre- mental sliding mode control by Dong et al. and a blend of fuzzy control and adaptive control [128–131]. Some representative methods are adaptive sliding mode fuzzy control by Liu et al. [145] and Xi and Hesketh [146]. [139–141].e. gain scheduling [132]. [149. combination of fuzzy control and LQR design [120–123]. [144]. These recent contributions can be divided into two types. integral sliding mode control by Defoort et al. terminal sliding mode control by Cao et al. [151]. and hierarchical sliding mode control by Tuan and Lee [147] to name but a few. the second type can also be subcategorized as fusion of fuzzy logic and evolutionary algorithms [118. second-order sliding mode control by Bartolini et al. One type can be categorized as fuzzy modeling [114–117].2. [152. most reports have focused on combination of fuzzy logic and other control techniques. The other type can be categorized as fusion of multiple control methods. Pieper and Surgenor [135] first explored the sliding mode control method for gantry cranes. [76] and Chang et al. i. fuzzy sliding mode control by Liu et al. [155]. Since then. which is abundant because of the diversity of control methods. the versatile sliding mode control method has received considerable devotion from researchers. [137. [143]. 1. inference mechanism of fuzzy systems [111]. The first type includes discrete sliding mode control by Pieper and Surgenor [136]. 110]. 153]. 138] and Vazquez et al. More recent reports about fusion of fuzzy control of other control methods. to name but a few. some investigations about crane control problems by sliding mode have been reported in recent years. where one of the control method is fuzzy logic. utilization of fuzzy control and proportional-integral- differential control [124–127]. neural network sliding mode control by Wang et al.7 Sliding Mode Control Owing to the nature of strong robustness. [148]. [154] and Tsai et al. suboptimal integral sliding mode control by Liu et al. 119]. 150] and Sun et al. The second type is characterized by a blend of sliding mode control and other control methods.32 1 Introduction design of fuzzy rules [106–108]..

Although the methodology of sliding mode control is of invariance against matched uncertainties. Crane systems inherently have highly nonlinear characteristics. etc. flexible cables. fuzzy sliding mode control [148–151]. un- modelling errors. 159] and incremental sliding mode control [143]. optimal discrete sliding mode control [136]. Concerning the details of the aforementioned literature. Some reports on this aspect have been reported. i. wind disturbances. This section reviews these issues from the points of view of theoretical and practical challenges. 1. Optimize the control command generated by sliding mode. optimal integral sliding mode control [152].e. adaptive sliding mode control [74. Design some novel structures of sliding surfaces. a blend of sliding mode control and other control methods are reported. 1.3 A Review of Crane Control 33 The sliding mode-based crane control has grown exponentially in the past two decades. Although the dynamics of cranes has been well under- stood. To improve the control performance of sliding mode control systems. All the practical factors are ideal during the crane modeling. 155]. 156–160]. How to suppress the effects of unmatched uncertainties on the stability of sliding mode control systems becomes crucial and important. . i. which may impede direct applications of mature sliding mode design. not all of these control methods for crane systems are practically applicable because of system constraints (e. The aspect is exhibited by applica- tions of mature sliding mode design for cranes to verify the feasibility. for instance. the factors cover friction nonlinearity.e. and observer-based sliding mode control [161]. Vazquez et al. their contributions mainly focus on the following viewpoints: Apply the method of sliding mode control. not all the uncertainties are as ideal as matched uncertainties..g.4 Challenges of Sliding Mode-Based Crane Control Owing to inherent nonlinearities. and external disturbances. parameter changes.1. hierarchical sliding mode control [147. To reject the factors.. and suboptimal sliding mode control [153].1 Theoretical Challenges Unmatched uncertainty.. practical cranes cover uncertainties. for example. Most of the early investigations are in the aspect of. Suppress uncertainties in crane dynamics. [139–141] and Xi and Hesketh [146] explored this field. For the purpose of illustration. sliding mode control is sometime compound with optimization methods. payload mass fluctuations. There are many uncertain factors in crane systems. actuator power lim- itation). neural network sliding mode control [154.4. for example. discrete sliding mode control [136] and second-order sliding mode control [137. 138].

However.. trolley position and its velocity are easy to measure but payload angle and angular velocity are hard to measure. Guaranteed Stability. there is always a desiderative expectation of new applications of crane systems that requires autonomous operation in an unstructured and pos- sibly dynamically changing environment. communication package delay or loss may lead to poor control performance. predictive control) is required. other sliding mode design methods in the field are kept untouched and remain problematic. Although crane sys- tems have received many well-established successes in well-structured environ- ment. Another practical concern is networked control. However. Fault tolerant detection and control. First. the issue is to be considered by scientists and engineers in the design of crane control system and a fault detection mechanism is highly recom- mended [161]. since the communication between actuator and sensor can be interrupted [58]. respectively. 1. Once network is introduced. indicating that crane control systems have a control input saturation constraint.g. cranes move on rails that are subject to limited lengths and payload angles should have their limits. Crane systems are widely used in industry. cranes are driven by actuators that are subject to their rated powers. Third. . Sensor failure results in delayed or missing feedback signal which affects overall tracking performance of crane control sys- tems. So far. not all variables of crane systems are measurable. the stability analysis is rather difficult under those ideas that directly optimize or plan the control command generated by sliding mode control [152. this fact hints the mea- surability constraints of crane systems. and a network control technique (e. Crane control systems should be robust to network delay. Usually. 153].4. a blend of sliding mode control and optimization methods can complicate the system stability analysis. Therefore. Networked crane control systems.34 1 Introduction They analyzed the effects of unmatched uncertainties on the stabilities of the dis- crete sliding mode and the second-order sliding mode closed-loop control systems. meaning that variables of crane systems have their boundary constraints. Especially. their control techniques for complex uncertain environment are still immature. Physical constraints. Second. A sole sliding mode control system has guaranteed sta- bility.2 Practical Challenges Industrial needs.

'interpreter'.y2 plot(yval(:.5*x1'.y2 plot(yval(:.'x1(0)=1'.sol. tval=(0:0.2 of yval vs.'latex'). .5b Plot program: figure(1). lgh=ylabel(latexStr). latexStr=['$$\dot x$$'] . latexStr=['$$\dot x$$'] .'x1(0)=1'. end y1(i+1)=y2(i)*dt+y1(i).5.yval(:.-1 1]).'t').-1 1]).'k') % plot col. tval=(0:0.xlabel('x').'k'). % column vector with t-values yval=double(subs([sol.'k') % plot col.y2.1 of yval axis([-1.'interpreter'. 1. 1. col.x2].'x2(0)=0'. B Matlab Codes to Plot Fig. set(lgh.Appendices 35 Appendices A Matlab Codes to Plot Fig.5*x1'.5.tval)).1). xlabel('x').sol.'t'). xlabel('x').1.'Dx2=-0.001. lgh=ylabel(latexStr).'x2(0)=0'.'latex'). dt=0.5a Plot program: figure(1).1).'interpreter'.1:10)'.2). set(lgh.y2(1)=0.'latex').1 of yval axis([-1. latexStr=['$$\dot x$$'] .x2].1:10)'.x1.tval)). for i=1:1:30000 if((y1(i)*y2(i))<0) k(i)=0. else k(i)=5.yval(:.1. % 2 columns with y1. end plot(y1. 1. col.'Dx2=-5. set(lgh. % column vector with t-values yval=double(subs([sol.2). C Matlab Codes to Plot Fig.2 of yval vs. % 2 columns with y1. y2(i+1)=y2(i)-dt*k(i)*y1(i). sol=dsolve('Dx1=x2'.'t'. sol=dsolve('Dx1=x2'.'t'.x1.7 Plot program: figure(1) y1(1)=1. lgh=ylabel(latexStr).

simout(:.2]) subplot(1.-1.').1).2.ylabel('x').ylabel('u').2.5.8 and 1. . figure(2)%% control signal subplot(1. 1.1].plot(t.simout(:. set(lgh. plot([0.'k').plot(t.'interpreter'.2).3). xlabel('t').xlabel('t').36 1 Introduction D Simulink Model to Plot Figs.'k').-1].1.hold on.1).1)).'k-.[0.1].hold on.2.-1].m figure(1)%%phase portrait of m=1 plot(simout(:.[0. lgh=ylabel(latexStr). xlabel('x').axis([-0.'latex'). latexStr=['$$\dot x$$'] .[0.2).simout(:.9 Plot program: signSMC_plot_Fig1_1819.1.

lgh=ylabel(latexStr).2.1)).1). set(lgh.11 Plot program: equSMC_plot_Fig1_1011.1).simout(:.'k-.m figure(1)%%phase portrait of m=1&k=1 plot(simout(:. latexStr=['$$\dot x$$'] .xlabel('t').ylabel('x').-1].'k').plot(t. xlabel('t'). axis([-0.-1.').2).'latex').[0.1.[0.5. figure(2)%% control signal subplot(1. xlabel('x').2]) subplot(1. .-1]. 1.'interpreter'.1].simout(:. ylabel('u').'k').plot(t.hold on.10 and 1.3).2. plot([0.[0.hold on.2).1.1].2.simout(:.Appendices 37 E Simulink Model to Plot Figs.

5). y3=x.2'. y1=sign(x).'k-.') hold on.'deta=0. plot(x.y3./(abs(x)+1)./(abs(x)+0.2.14 .y2.-1.'deta=0. y2=x.y4. y4=x.m x=-100:1:100. plot(x. 1. axis([-110.13 and 1.'deta=1') G Simulink Model to Plot Figs.'k:') hold on.110.38 1 Introduction F Matlab Codes to Plot Fig. plot(x. 1.12 Plot program: Fig1_12. plot(x.2).y1.5'./(abs(x)+0.2]) legend('no deta'.1.'k--').'k-') hold on.

1].1 plot(simout(:.16 .ylabel('x').simout(:.1]. xlabel('x').-1]. lgh=ylabel(latexStr).'k-.'k').Appendices 39 Plot program: deltaSMC_plot_Fig1_1314 figure(1)%%phase portrait of m=1 & k=1 & deta=0.3). set(lgh.10.1. subplot(1.'latex').plot(t. hold on.plot(t.'interpreter'.2]).2.15 and 1.1.xlabel('t').1).ylabel('u'). figure(2)%% control signal subplot(1.-1]. H Simulink Model to Plot Figs. axis([-0.').simout(:. latexStr=['$$\dot x$$'] .-1. hold on.2).2).[0.[0.2.1)).1). plot([0.'k').xlabel('t').simout(:.[0. 1.2.

NumOutputs = 2.x.1). function sys=mdlDerivatives(t.simout(:.'interpreter'. m function [sys.[0. sizes.plot(t.ts]=mdlInitializeSizes sizes = simsizes. figure(2)%% control signal subplot(1. latexStr=['$$\dot x$$'] .x.1].NumInputs = 1. sys(2)=x(2).'k-.3. sizes. sys=mdlOutputs(t.NumSampleTimes = 0. ts = [].x. subplot(1. sizes. sys = simsizes(sizes).2). .plot(t.'k'). sys(2)=u(1)+0.1). end function [sys. sys=mdlDerivatives(t.ylabel('u'). plot([0.NumContStates = 2. subplot(1.simout(:.m figure(1)%%phase portrait of m=1&k=1 plot(simout(:. case 3.'latex').str.ts] = spacemodel(t.4)) xlabel('t').3.u).9} sys=[]. lgh=ylabel(latexStr).x0.[0. x0 = [1 0].DirFeedthrough = 0.u).-1].x0.plot(t.ylabel('s').num2str(flag)]).3).ts]=mdlInitializeSizes.str.NumDiscStates = 0.hold on.-1].ylabel('x').u) sys(1)=x(1).1*rand().flag) switch flag. str = [].4.'k'). otherwise error(['Unhandled flag = '. function sys=mdlOutputs(t. case 1.40 1 Introduction Plant program: plant. Plot program: equSMC_plot_Fig1_1516. xlabel('x').simout(:.1].3.x0.u) sys(1)=x(2).hold on.[0.1)) xlabel('t'). sizes. xlabel('t').x. case {2. set(lgh. case 0. sizes.').u.str.3). sizes.simout(:.x.2). [sys.

sizes.u. str = []. case 1. otherwise error(['Unhandled flag = '.x0. 1.9} sys=[].ts]=mdlInitializeSizes sizes = simsizes.20 Plant program: plant. sys=mdlOutputs(t. sizes. sys = simsizes(sizes). x0 = [1 0]. ts = []. sizes. sys=mdlDerivatives(t.x0.str.18. [sys.str.u).x. sizes.Appendices 41 I Simulink Model to Plot Figs.NumInputs = 1.NumSampleTimes = 0.19 and 1. sizes. case 0. case {2. case 3.x. .ts] = spacemodel(t. 1.4.DirFeedthrough = 0.x0.NumContStates = 2.flag) switch flag.x.str. end function [sys.NumDiscStates = 0.ts]=mdlInitializeSizes.num2str(flag)]).17.NumOutputs = 2.u). sizes. m function [sys. 1.

Optim Control Appl Meth 3(2):115–120 11.plot(t. IEEE Trans Control Syst Technol 18(6):1345–1358 5.u) sys(1)=x(1). Homaifar A.1].simout(:. Springer. New Jersey 7.1*rand().2).1].'k'). DC. Georgia Institute of Technology 4. figure(2)%% control signal subplot(1. Kiniaghalam B. Washington.3. Grizzle JW (1996) Nonlinear systems. design. function sys=mdlOutputs(t. pp 2124–2130 12.[0. latexStr=['$$\dot x$$'] .hold on. sys(2)=u(1). References 1. Prentice Hall.3. Berlin 6. Vaughan J. subplot(1. Field JA (1961) The optimization of the performance of an ore bridge. Dissertation.42 1 Introduction function sys=mdlDerivatives(t.ylabel('x'). and simulation. Lancaster L (1999) Building Trajan’s column. J Dyn Syst Meas Control Trans ASME 126(2):359–364 .4)) xlabel('t').simout(:.1)) xlabel('t'). Am J Archaeol 103(3):419–439 3. Trans Eng Inst Canada 5(3):163–169 10. Padstow 8.[0. plot([0. xlabel('x'). Bikdash M.ylabel('s') subplot(1. Vaughan J (2008) Dynamics and control of mobile cranes. Edwards C.x. Kurrer KE (2008) The history of the theory of structures: from arch analysis to computational mechanics. Plot program: equSMC_plot_Fig1_17_20.3). 2nd edn. Utikin VI (1992) Sliding modes in control and optimization.sys(2)=x(2). Spurgeon S (1998) Sliding mode control: theory and applications. New Jersey 9. USA.'). lgh=ylabel(latexStr).3).'k'). set(lgh.plot(t. CRC Press. Khalil HK.'latex').'k-.-1].plot(t. Kim D. Ernst & Sohn.[0.'interpreter'. xlabel('t').2).1).m figure(1)%%phase portrait of m=1&k=1 plot(simout(:.-1]. Dozier G (1999) Genetic algorithms solution for unconstrained optimal crane control.simout(:. Prentice Hall. Lee HH (2004) A new motion-planning scheme for overhead cranes with high-speed hoisting.simout(:. Manson GA (1982) Time-optimal control of an overhead crane model. Berlin 2.hold on. In: Proceedings of the congress on evolutionary computation. Bequette BW (2003) Process control: modeling.u) sys(1)=x(2)+0.1).3.x.ylabel('u'). Singhose W (2010) Control of tower cranes with double-pendulum payload dynamics.

Khalid A. Miranda E (2008) Computational methods for coordinating multiple construction cranes. Dohrmann C. IEEE ASME Trans Mechatron 17(1):166–173 22. doi:10. USA. Sorensen K. In: Proceedings of the 1999 IEEE international conference on control applications. J Dyn Syst Meas Control Trans ASME 128(4):842– 845 15. Seering W (1999) Vibration reduction with specified-swing input shapers. J Dyn Syst Meas Control Trans ASME 121(2):242–247 27. Tokyo. IET Control Theory Appl 6 (10):1410–1423 23. Seering W (1997) Convolved and simultaneous two-mode input shapers. Hammad A (2012) Near real-time motion planning and simulation of cranes in construction: framework and system architecture. Dickerson S (2007) A controller enabling precise positioning and sway reduction in bridge and gantry cranes. Seering W (1999) An expert system for the design of input shapers. Mills B.2936384 19. Zameroski D. Fujikawa H. USA. Kolodziejczyk K (2007) Motion planning and control of gantry cranes in cluttered work environment. Lawrence J. Ross IM. USA. Optim Control Appl Meth 34(2):191–201 20. Augusta. J Dyn Syst Meas Control Trans ASME 128 (4):835–841 28. In: Proceedings of American control conference. Frakes D (2006) Human operator performance testing using an input-shaped bridge crane. Parker GG. Starr G. Blajer W. Wood J. Int J Control 78(12):875–886 14. Huey J. pp 934–938 26. GA. Forster GA. Zhang X. Yuan Y (2012) Transportation task-oriented trajectory planning for underactuated overhead cranes using geometric analysis. Da Cruz JJ. Liang Y. Singer N. Seattle. Lumia R (2008) Rapid swing-free transport of non-linear payloads using dynamic programming. San Francisco. Jones S. In: Proceedings of IFAC workshop on systems engineering for control systems. Matsumoto K (1983) Suboptimal control of the roof crane by using the microcomputer. Kohala Coast-Island. Kriikku E (1997) An input shaping controller enabling cranes to move without sway. Lin TC (1993) Design an input shaper to reduce operation-induced vibration. Singhose W. Daqaq MF. Nonlinear Dyn 45(1–2):149–170 .References 43 13. Alsop CF. IET Control Theory Appl 1(5):1370–1379 17. In: Proceedings of the 1999 IEEE international conference on control applications. Singhose W. Leonardi F (2013) Minimum-time anti-swing motion planning of cranes using linear programming. USA. Segura D (2006) A sliding-mode antiswing trajectory control for overhead cranes with high-speed load hoisting. pp 323–328 30. Sun N. Fang YC. pp 295–305 29. Ulsoy AG (1999) An approach to control input shaping with application to coordinate measuring machines. pp 533–538 34. Kohala Coast-Island. San Francisco. Singhose W. Zhang YD. Petterson B. pp 2502–2506 31. Kang SC. Holmes ER (1965) Ore unloader automation-A feasibility study. J Comput Civ Eng 22(4):252–263 18. IEE Proc Control Theory Appl 144(6):515–520 32. Singhose W. French L. In: Proceedings of the annual conference on industrial electronics. Crain E. J Comput Civ Eng 26(1):54–63 21. IEEE Trans Autom Control 49(8):1410–1413 16. Japan. J Dyn Syst Meas Control Trans ASME 130(4). Sun N. Singhose W. Control Eng Pract 15(7):825–837 25. In: Proceedings of 7th topical meeting on robotics and remote systems. pp 713–718 33. Lee HH (2005) Motion planning for three-dimensional overhead cranes with high-speed load hoisting. Fang Y. Ma BJ (2012) A novel kinematic coupling-based trajectory planning method for overhead cranes. pp 225–231 24. Singhouse W. Yamada S. Masoud ZN (2006) Nonlinear input-shaping controller for quay-side container cranes. USA. In: Proceedings of American control conference. Fahroo F (2004) Pseudospectral methods for optimal motion planning of differentially flat systems. Lee HH. Robinett RD (1995) Command shaping for residual vibration free crane maneuvers. AlBahnassi H.1115/1.

IEEE Trans Autom Control 45(8):1498–1502 51. In: Proceedings of IEEE international conference on automation science and engineering. He XX. In: Proceedings of international conference on control and automation. France. In: Proceedings of international conference on control. Yoon JS. pp 472–479 43. Lee SG (2011) Feedback linearization control of overhead cranes with varying cable length. Kim GH. Paris. Monterey. Masoud ZN. Munich. J Electr Electron Eng Aust 12:75–84 42. In: Proceedings of IEEE international conference on robotics and automation.44 1 Introduction 35. pp 546–549 55. Wu XQ. Thull D. Germany. pp 1198–1203 48. Chiang Mai. Nguyen HT (1994) State-variable feedback controller for an overhead crane. Korea. Thailand. IEEE Trans Ind Electron 59 (12):4723–4734 53. Ortega R. Caron B. pp 535–542 41. pp 4473–4478 49. Sung YG. Virkkunen J (1995) Optimal path planning for a trolley crane: fast and smooth transfer of load. Seoul. Sun N. Wild D. pp 2196–2201 40. Marttinen A. In: Proceedings of IEEE international conference on control applications. Hungary. IEE Proc Control Theory Appl 142(1):51–57 44. Karkoub MA. Grassin N. Lee JS. J Vib Control 5(3):443–459 54. Yi J. USA. In: Abstracts of the IFAC 5th world congress. Grenoble. Guo W. Park BS. Retz T. Maschke B. Liu D. Karlsruhe. pp 906–911 47. Fang YC (2012) New energy analytical results for the regulation of underactuated overhead cranes: an end-effector motion-based approach. Fantoni I (2000) Control of convey-crane based on passivity. Korea. Baharova L. Janchiv A. Tuan L. Kim GH. In: Proceedings of the 1st European control conference. pp 2821–2826 50. Massachusetts Institute of Technology 45. USA. Bourles H et al (1991) Robust control of a travelling crane. In: Proceedings of the ASME dynamic systems and control division. Masoud Z. Chen H (2005) Nonlinear controller for a gantry crane based on partial feedback linearization. Gyeonggi-do. IEEE Trans Control Syst Technol 14(6):1070–1077 36. Lee SG (2012) Partial feedback linearization control of the three dimensional overhead crane. Fang YC (2014) A novel anti-swing control method for 3-D overhead cranes. Portland OR. Singh I (1999) Passive control of overhead cranes. Lee HH (1997) Modelling and control of 2-dimensional overhead crane. Dallas. France. Van der Schaft AJ (2000) Energy-based Lyapunov functions for forced Hamiltonian systems with dissipation. 1115/1. Zhang X. IEE Proc Control Theory Appl 149(3):209–216 52. Hazlerigg A (1972) Automatic control of crane operations. Singhose W (2009) Limited-state commands for systems with two flexible modes.and passivity-based control concept to a crane with heavy chains and payload. Park HS (1995) Various control schemes for implementation of the anti-swing crane. pp 11–13 39. Dissertation. Sun N. Olfati-Saber R (2001) Nonlinear control of underactuated mechanical systems with application to robotics and aerospace vehicles. In: Proceedings of American control conference. Sun N. Lozano R. In: Proceedings of the ANS 6th topical meeting on robotics and remote systems. USA. Alhazza K (2014) Frequency-modulation input shaping control of double-pendulum overhead cranes. Zribi M (2002) Modelling and energy based nonlinear control of crane lifters.4025796 38. Collado J. Zhao D (2004) Passivity-based-control for double-pendulum-type overhead cranes. USA. J Dyn Syst Meas Control Trans ASME 136(2). pp 656–661 . In: Proceedings of American control conference. Kugi A (2006) Application of a combined flatness. doi:10. automation and systems. Alli H. Mechatronics 19(5):780–787 37. In: Proceedings of IEEE region 10 annual international conference. Tuan L. Gao B. Germany. pp 1074–1078 46. Hamalainen JJ. Fang YC (2013) A partially saturated nonlinear controller for overhead cranes with experimental implementation. pp 1260–1264 56. Chicago IL. Budapest. Daqaq MF (2006) A graphical approach to input-shaping control design for container cranes with hoist.

Chelmsford 66. Nice. Desantis RM (1983) Microprocessor-based adaptive control of a crane system. Zhang S. Zhou KM. USA. pp 1832–1837 71. In: Proceedings of the 33rd Chinese control conference. Tsai CC. He W. Krstic M. Fang Y. In: Proceedings of the 30th IEEE conference on decision and control. Liu D. Kokotovic PV (1992) The joy of feedback: nonlinear and adaptive. Seattle WA. Dixon WE. Liu Y. IEEE Control Syst Mag 11(1):57–62 68. In: Proceedings of IEEE/ASME international conference on advanced intelligent mechatronics. Wang P. pp 1963–1968 70. In: Proceedings of 1992 IEEE international conference on robotics and automation. Boustany F (1991) Adaptive control of a class of mechanical systems using linearization and Lyapunov methods: a comparative study on the overhead crane example. In: Proceedings of the 31st IEEE conference on decision and control. Wang HW. Ge SS (2014) Adaptive control of a flexible crane system with the boundary output constraint. Chen H (2014) Adaptive control of underactuated crane systems subject to bridge length limitation and parametric uncertainties. Minneapolis MN. Zhao D. Dawson DM. IEEE Control Syst Mag 12(3):7–17 60. Wang W (2005) Adaptive sliding mode fuzzy control for a two-dimensional overhead crane. In: Proceedings of the 22nd IEEE conference on decision and control. Akita. Prentice Hall. Taiwan. Serrani A (2008) Offshore crane control based on adaptive external models. pp 3568–3573 73. In: Proceedings of American control conference. Yang JH. Messineo S. Wu HL. USA. IEEE ASME Trans Mechatron 8(3):418–423 58. Åström KJ. Butler H. Courier Corporation. In: Proceedings of 51st annual conference of the society of instrument and control engineers of Japan. d’Andrea-Novel B (1992) Adaptive control of an overhead crane using dynamic feedback linearization and estimation design. Ngo QH. IET Control Theory Appl 7(7):921–935 59. Qingdao. Rodriguez FO (2011) Stable adaptive compensation with fuzzy CMAC for an overhead crane. Sun N. Kaohsiung. Fang Y. France. Brighton. pp 1256–1261 65. Kokotovic PV. Kanellakopoulos I (1992) Backstepping to passivity: recursive design of adaptive systems. Automatica 36(4):587–593 62. Zhang X (2012) A motion planning-based adaptive control method for an underactuated crane system. Wittenmark B (2013) Adaptive control. Japan. pp 944–947 67. USA. Chuang KH (2012) Backstepping aggregated sliding-mode motion control for automatic 3D overhead cranes. Yang JH. Hurteau R. Tucson AZ. Boustany F. New Jersey . d’ Andréa-Novel B. pp 120–125 69. Yang KS (2007) Adaptive coupling control for overhead crane systems. IEEE Trans Control Syst Technol 20(1):241–248 76. Zergeroglu E (2003) Nonlinear coupling control laws for an underactuated overhead crane system. Nanjing. Yang KS (2006) Adaptive control for 3-D overhead crane systems. In: Proceedings of American control conference. Moreno-Armendariz MA. Cao LZ. pp 3276–3280 61. China.References 45 57. Hong KS (2012) Adaptive sliding mode control of container cranes. van Amerongen J (1991) Model reference adaptive control of a gantry crane scale model. IET Control Theory Appl 6(5):662–668 75. England. pp 849–854 64. pp 2498–2503 72. Wu HL. Chuang KH (2012) Intelligent sliding-mode motion control using fuzzy wavelet networks for automatic 3D overhead cranes. Fang Y. IEEE Trans Ind Electron 61(8):4126–4133 74. Mechatronics 17(2):143–152 78. Yi J. Inf Sci 181(21):4895–4907 79. Doyle CJ (1999) Essentials of robust control. USA. In: Proceedings of IEEE international conference on automation and logistics. Mechatronics 15(5):505–522 77. San Antonio TX. China. Ma B. Wei SB (2007) Adaptive backstepping control of crane hoisting system. Niu C. pp 245–248 63. Yu W. Coron JM (2000) Exponential stabilization of an overhead crane with flexible cable via a back-stepping approach. Yu HN (2013) A survey of underactuated mechanical systems. Honderd G. d’Andrea-Novel B. Tsai CC.

In: Proceedings of 2012 IEEE international conference on industrial technology. pp 846–851 90. Deng MC. In: Proceedings of American control conference. Becerra VM (2004) Application of constrained predictive control on a 3D crane system. Neupert J. Sawodny O. pp 647–652 84. Arnold E. Sawodny O (2014) Observer design and flat-ness-based feedforward control with model predictive trajectory planning of a crane rotator. Bartulovic M. IEEE ASME Trans Mechatron 5(1):23–31 82. Uchiyama N. Wang DY (2011) Operator-based robust nonlinear control for a crane system with constraint inputs. pp 6109–6114 88. In: Proceedings of IEEE conference on robotics. USA. Uchiyama N (2009) Robust control of rotary crane by partial-state feedback with integrator. pp 583–587 94. pp 1533–1538 95. Int J Model Ident Control 16(1):86–96 89. Sayarrodsari. Mechatronics 19(8):1294–1302 86. pp 3973–3978 97.46 1 Introduction 80. Homaifar A. Australia. Matusko J. Seto K. Busan. Uchiyama N (2012) Robust control of rotary cranes under rope length variance via LMI optimization. pp 785–790 91. Schaper U. Ushida Y. Takami I (2014) Robust control for crane considering all varying parameters in the dynamics. Bauer D. Portland OR. Deng JM. Kimiaghalam B. Yantai. Arnold E. pp 680–685 98. Inoue A (2012) Operator-based robust non-linear control for gantry crane system with soft measurement of swing angle. Istanbul. Kapernick B. Wang DY (2011) Operator-based robust right coprime factorization design for planar gantry crane. In: Proceedings of the 16th . Ohno Y. In: Proceedings of IEEE international conference on mechatronics and automation. Doi F (2000) Multimode vibration control of a flexible structure using H∞- based robust control. automation and mechatronics. Schneider K (2013) Constrained real-time model-predictive reference trajectory planning for rotary cranes. Bordons C (2004) Model predictive control. IEEE Trans Autom Control 25(6):1058–1072 81. Washington DC. Toda M (2007) A unified approach to robust control of flexible mechanical systems. Narita M. Swevers J (2011) A model predictive control approach for time optimal point-to-point motion control. Berlin 92. B (2001) An application of model predictive control for a shipboard crane. pp 5787–5793 85. China. pp 2724–2729 83. In: Proceedings of American control conference. Wen SJ. Schneider K. In: Proceedings of 2006 IEEE/RSJ international conference on intelligent robots and systems. Taichung. USA. Wollongong. Beijing. Diehl M. pp 1–5 87. Yang TW. Camacho EF. In: Proceedings of the 9th IEEE international workshop on advanced motion control. pp 4020–4025 99. Chen G. In: Proceedings of 2010 international conference on mechatronics and automation. Sano S (2006) Robust control of rotary cranes based on pole placement approach. Sano S. New Orleans LA. Wen SJ. Schneider K (2005) Anti-sway system for boom cranes based on a model predictive control approach. Wen SJ. Van den Broeck L. Sawodny O. Graichen K (2013) Model predictive control of an overhead crane using constraint substitution. Springer. Ouyang H. Arlington VA. In: Proceedings of the 11th IEEE international conference on control and automation. Deng MC. USA. Deng MD. Ushiro T. USA. China. Xi’an. Kar IN. In: Proceedings of American control conference. Zuzic G. Singapore. Turkey. Iles S. Taiwan. Mechatronics 21(7):1203–1212 96. Takagi S. In: Proceedings of the 46th IEEE conference on decision and control. O’Connor WJ (2006) Wave based robust control of a crane system. Barisa T. China. In: Proceedings of the 30th Chinese control conference. Kolonic F (2014) Nonlinear predictive control of a tower crane using reference shaping approach. In: Proceedings of IEEE/ASME international conference on advanced intelligent mechatronics. Schaper U. Ackermann J (1980) Parameter space design of robust control systems. Korea. Niagara Falls. Canada. pp 929–934 93.

Antsaklis PJ. New York 106. Precup RE. Smoczek J. Takeuchi O. In: Proceedings of international conference on industrial electronics. Yubazaki N.References 47 international power electronics and motion control conference and exposition. Petrenko YN. Kruse R. Stengel RF (1991) Intelligent failure tolerant control. In: Proceedings of international conference on industrial electronics. pp 224–230 101. Itoh O. Filip HI. Yi J. Yamada S. Irie Y (1993) Application of fuzzy control to automatic crane operation. Wang WJ. Antalya. IEEE Control Syst Mag 11(4):14–23 104. China. control and power engineering. Wakasugi Y (1989) Fuzzy control of the roof crane. Yamada S. Fujikawa H (1993) Anti-swing control of the container crane by fuzzy control. Matusko J. and instrumentation. In: Proceedings of the 14th annual conference of IEEE industrial electronics society. pp 709–714 108. Numberger A (2006) Fuzzy control: fundamentals. Singapore. Tan M (2002) Proposal of GA-based two-stage fuzzy control of over-head crane. Yamada S. Benhidjeb A. and instrumentation. Fujikawa H (1995) An optimization of fuzzy controller and its application to overhead crane. pp 230–235 109. Khatamianfar A. Liu D. Control Eng Pract 3(12):1687–1696 114. Chen YJ. Passino KM (1993) An introduction to intelligent and autonomous control. Ohtsubo R. Turkey. pp 415–420 107. and instrumentation. Yamada S (1988) The application of fuzzy theory for a rotary crane control. USA. Appl Soft Comput J 24:1155–1163 118. In: Proceedings of the 2002 IEEE international conference on fuzzy systems. IEEE Trans Control Syst Technol 22(4):1461–1473 103. Russia. Itoh J. Hadipriono Tan F (2009) Mobile crane safe operation approach to prevent electrocution using fuzzy-set logic models. Japan. Springer. Kijima Y. Kolonic F (2014) Real-time predictive control of 3D tower crane. control. communications. USA. Yi J. Alavi SE (2010) Fuzzy logic and genetic algorithm technique for non-liner system of overhead crane. In: Proceedings of the 15th annual conference of IEEE industrial electronics society. Al-Humaidi HM. pp 848–851 . Beijing. Suzuki Y. Wozniak P (1996) On the robustness of fuzzy control of an overhead crane. Warsaw. Maui HI. pp 161–164 110. Fujikawa H. Migita H. USA. pp 1298–1303 112. Bock M. In: Proceedings of international conference on industrial electronics. Takeuchi S. In: Proceedings of 2002 IEEE region 10 conference on computer. Klawonn F. Owczarz D. Irkutsk Listvyanka. Fujikawa H. pp 872–876 100. Iles S. In: Proceedings of IEEE 23rd international symposium on industrial electronics. Kluwer Academic Publishers. pp 1508–1513 113. France. Istanbul. Dordrecht 105. In: Proceedings of 2010 IEEE region 8 international conference on computational technologies in electrical and electronics engineering. Eng Appl Artif Intell 28:190–200 117. control. Adv Eng Softw 40(8):686–696 115. Maui HI. Gissinger GL (1995) Fuzzy control of an overhead crane performance comparison with classic control. Kugi A (2014) Real-time nonlinear model predictive path-following control of a laboratory tower crane. Szpytko J (2014) Evolutionary algorithm-based design of a fuzzy TBF predictive model and TSK fuzzy anti-sway crane control system. stability and design of fuzzy controllers. Turkey. pp 433–437 111. Strasbourg. Chang CL (2009) Guaranteed cost control for an overhead crane with practical constraints: Fuzzy descriptor system approach. Orlando FL. Philadelphia PA. Hirota K (2002) Anti-swing fuzzy control of overhead traveling crane. In: Proceedings of European control conference. USA. USA. Savkin AV (2014) A new tracking control approach for 3D over-head crane systems using model predictive control. In: Proceedings of the IEEE international symposium on industrial electronics. Michels K. Rădac MB et al (2014) Online identification of evolving Takagi– Sugeno–Kang fuzzy models for crane systems. Honolulu HI. Eng Appl Artif Intell 22:639–645 116. control. Nowacki Z. pp 1721–1724 119. pp 796–801 102.

pp 573–578 133. pp 390–395 123. Malaysia. Yang Y (2008) Fuzzy-PI damping control for hydraulic crane tip. computing and engineering. computing and engineering. El-Awady H et al (2014) Control of overhead crane based on Lyapunov approach to fuzzy controller synthesis. IEEE Trans Ind Inf 3(2):164–172 129. pp 92–97 132. Szpytko J (2014) Iterative and evolutionary optimization for interval analysis-based designing a fuzzy controller for a planar crane model. pp 1–6 134. In: Proceedings of 17th international conference on methods and models in automation and robotics. China. Międzyzdroje. pp 306–311 122. Zhai X. Adeli M. Surgenor BW (1994) Discrete time sliding mode control applied to a gantry crane. Adeli M. Zarabadipour H. Lake Buena Vista FL. Chwa D. Cui Y (2014) Research on application of fuzzy adaptive PID controller in bridge crane control system. Seoul. Kuala Lumpur. pp 829–834 136. Shoorehdeli MA (2011) Anti-swing control of a double-pendulum-type overhead crane using parallel distributed fuzzy LQR controller. Zhao H. Lin F. automation and systems. In: Proceedings of IEEE international conference on advanced communication control and computing technologies. Malaysia.48 1 Introduction 120. Zarabadi SH. pp 401–406 124. pp 2886–2891 130. Mudi RK (2012) An adaptive fuzzy controller for overhead crane. Lake Buena Vista FL. In: Proceedings of 18th international conference on methods and models in automation and robotics. pp 1092–1097 125. Smolenice Castle. In: Proceedings of IEEE international conference on software engineering and service science. In: Proceedings of international conference on control. South Korea. Aliyari Shoorehdeli M (2011) Design of a parallel distributed fuzzy LQR controller for double-pendulum-type overhead cranes. pp 300–304 131. Chang CY (2007) Adaptive fuzzy controller of the overhead cranes with nonlinear disturbance. USA. pp 62–67 121. Zarabadi SH. pp 829–834 . Adeli M. In: Proceedings of 19th international conference on methods and models in automation and robotics. Smoczek J. Zarabadipour H. Liu C. pp 971–974 128. Shoorehdeli MA (2011) Crane control via parallel distributed fuzzy LQR controller using genetic fuzzy rule selection. Penang. Zhang X. Malaysia. Solihin MI. pp 258–263 135. Ramanathapuram. USA. Adeli M. In: Proceedings of 2007 international conference on intelligent and advanced systems. Huang H (2011) The application study of heave compensation control based on motion prediction and fuzzy-PID for intelligence crane. Penang. In: Proceeding of international conference on control. Zarabadipour H. Pieper JK. In: Proceedings of 33rd IEEE conference on decision and control. Beijing. Poland. In: Proceedings of IEEE international conference on control system. Park MS. China. Smoczek J. Iran. In: Proceedings of IEEE international conference on control system. Poland. Smoczek J. pp 75–79 126. Szpytko J (2013) Fuzzy logic-based adaptive control system prototypying for laboratory scaled overhead crane. In: Proceedings of 23rd international conference on robotics in Alpe-Adria-Danube region. Zarabadipour H. China. India. In: Proceedings of 33rd IEEE conference on decision and control. Szpytko J (2012) Design of gain scheduling anti-sway crane controller using genetic fuzzy system. Moustafa K. Jinan. Iran. instrumentation and automation. Shiraz. Shiraz. Slovakia. In: Proceedings of 5th international conference on fuzzy systems and knowledge discovery. Międzyzdroje. Pal AK. Beijing. Aliyari Shoorehdeli M (2011) Anti-swing control for a double-pendulum-type overhead crane via parallel distributed fuzzy LQR controller combined with genetic fuzzy rule set selection. Hong SK (2007) Adaptive fuzzy nonlinear anti-sway trajectory tracking control of uncertain overhead cranes with high-speed load hoisting motion. Sharkawy AB. In: Proceedings of international conference on mechatronics and automation. Pieper JK. In: Proceedings of international conference on control. Wahyudi (2007) Fuzzy-tuned PID control design for automatic gantry crane. instrumentation and automation. pp 2370–2374 127. Surgenor BW (1994) Discrete time sliding mode control applied to a gantry crane.

Florence. Lou FP. Freidovich L. Wang W (2004) Swing-free transporting of two-dimensional overhead crane using sliding mode fuzzy control. pp 2876–2880 141. Zhao DB. Orani N. Usai E (2000) Load swing damping in overhead cranes by sliding mode technique. pp 2541–2546 150. pp 484–487 145. Wang Z. Taipei. pp 2195–2200 153. Xi’an. China. Vazquez C. Liu R. Huang GE (2006) An enhanced adaptive sliding mode fuzzy control for positioning and anti-swing control of the overhead crane system. Cheng J (2008) Incremental sliding mode control for double-pendulum-type overhead crane system. In: Proceedings of international conference on intelligent mechatronics and automation. China. Liu DT. In: Proceedings of 24th Chinese control and decision conference. Harbin. Dong Y. pp 368–371 144. man and cybernetics. Chen WT. Aranovskiy S. Chang CY. Australia. Meng WJ (2012) Fuzzy sliding mode anti-swing control for tower crane base on time-delayed filter. Feng Z. In: Proceedings of 27th Chinese control conference. Defoort M. et al (2010) Controller design and simulation of the crane based on non-singular terminal sliding mode method. China. pp 1697–1702 138. Li S (2014) Suboptimal integral sliding mode controller design for a class of affine systems. J Mech Sci Technol 27(6):1863–1873 148. Chen ZM. Italy. Collado J (2013) Second order sliding mode control of an overhead-crane in the presence of external perturbations. pp 992–997 149. Fridman L. Tuan LA. Yi JQ. Taiwan. Vazquez C. In: Proceedings of IEEE international conference on systems. Los Angeles CA. In: Proceedings of American control conference. Liu DT. Canada. In: Proceedings of international conference on logistics systems and intelligent management. USA. China. Saif M (2008) Output feedback controller design for a class of MIMO nonlinear systems using high-order sliding-mode differentiators with application to a laboratory 3-D crane. Hesketh T (2010) Discrete time integral sliding mode control for overhead crane with uncertainties. IET Control Theory Appl 4(10):2071–2081 147. Chen X (2013) An optimal integral sliding mode control design based on pseudospectral method for overhead crane systems: In: Proceedings of 32nd Chinese control conference. In: Proceedings of 39th IEEE conference on decision and control. Xi’an. Fridman L (2014) Second order sliding mode control of a mobile hydraulic crane. Yi JQ. USA. Yi JQ. Xi Z. Pisano A. France. Fridman L (2012) On the second order sliding mode control of a parametrically excited overhead-crane. Pisano A. Wang W. pp 2205–2210 152. Li S. Chiang KH. Taiyuan. Lee SG (2013) Sliding mode controls of double-pendulum crane systems. Liu R. IEEE Trans Ind Electron 55(11):3985–3997 143. Kunming. Automatica 38(10):1783–1790 139. pp 71–77 146. Boston MA. pp 6288–6293 140. Zhao DB. In: Proceedings of IEEE 53rd annual conference on decision and control. Bartolini G. pp 5530–5535 142. Sydney. Liu XJ (2004) Incremental neural network sliding mode controller for an overhead crane. Vazquez C. Zhao DB (2003) Fuzzy tuning sliding mode control of transporting for an overhead crane.References 49 137. pp 166–171 . J Optim Theory Appl 161(3):877–904 154. Maneeratanaporn J. Usai E (2002) Second-order sliding-mode control of container cranes. China. Sun H. In: Proceedings of IEEE 52nd annual conference on decision and control. Paris. pp 1764–1769 151. In: Proceedings of international conference on machine learning and cybernetics. Hsu KC. Li XB. Chengdu. Cao L. Murakami T (2012) Integral sliding mode anti-sway control of an underactuated overhead crane system. Montréal. China. Bartolini G. In: Proceedings of 9th France-Japan and 7th Europe-Asia congress on mechatronics and research and education in mechatronics. In: Proceedings of American control conference. Collado J.

Tsai CC. Kim DH. In: Proceedings of 3rd international conference on signals. circuits and systems. Dhahri S. control. Moon SC.50 1 Introduction 155. Park MS. Thailand. Medenine. Tong SW. Wu HL. IEEE Trans Fuzzy Syst 22(5):1262–1271 161. and intelligent systems. Appl Math Inform Sci 7(4):1359–1364 159. Chuang KH (2012) Intelligent sliding-mode motion control using fuzzy wavelet networks for automatic 3D overhead cranes. Int J Precis Eng Manuf 14(8):1329–1338 160. pp 1256–1261 156. Nho LC. Tuan LA. Japan. Lee SG (2013) Adaptive sliding mode control of overhead cranes with varying cable length. Eom M (2014) Adaptive sliding-mode antisway control of uncertain overhead cranes with high-speed hoisting motion. Lee SG (2012) Adaptive sliding mode control of three dimensional overhead cranes. Yi JQ (2013) Adaptive control based on incremental hierarchical sliding mode for overhead crane systems. Chwa D. Hmida FB. Akita. Gossa M (2009) Actuator fault reconstruction for linear uncertain systems using sliding mode observer. Tuan LA. Tuan LA. Tunisia. Lee WG. In: Proceedings of IEEE international conference on cyber technology in automation. Moon SC. pp 354–359 157. pp 1–6 . Qian DW. Kim DH (2013) Model reference adaptive sliding mode control for three dimensional overhead cranes. J Mech Sci Technol 27(3):885–893 158. Bangkok. In: Proceedings of 51st annual conference of the society of instrument and control engineers of Japan. Lee SG. Sellami A.

1 are described as the trolley mass M.Chapter 2 Crane Mathematic Model Abstract This chapter examines the dynamics of overhead cranes. trolley and payload [1]. the swing angle of the payload with respect to the vertical line θ.1.. 2. Yi. In addition. DOI 10. and the driven force applied to the trolley f.e. Qian and J.1 Modeling Figure 2. If the trolley moves toward the right direction by a positive driven force. Subsequently. 2. the trolley position with respect to the origin x.1 is static and the payload is in its downward position. i. The two models are pre- sented as references for examples throughout this book. Concerning single-pendulum-type overhead cranes. Other symbols in Fig.1 shows the coordinate system of an overhead crane system with its payload.1007/978-3-662-48417-3_2 . Apparently.1 Modeling of Single-Pendulum-Type Cranes 2. the payload angle θ is inherently © Springer-Verlag Berlin Heidelberg 2016 51 D. The analysis distills the essential properties of each. some uncertainties associated with real applications are discussed next. the chapter proceeds with the analysis of oscillations for pendulum-type motions on the basis of linearized models of the two types of overhead cranes. the payload mass m. the crane system consists of two subsystems. then the payload will rotate clockwise. Since the two models are established under some ideal assumptions. the rope length L. Consider that the crane in Fig. The latter is suspended from the trolley by a rope. their equations of motion are first presented by means of its Euler–Lagrange equations. Hierarchical Sliding Mode Control for Under-actuated Cranes. Keywords Overhead crane modeling  Single-pendulum dynamics   Double-pendulum dynamics Uncertainty 2. the equations of motion are extended to double-pendulum-type overhead cranes. Apparently. The former is driven by a force.

For the purpose of simplification. • No friction exists in the system. q1 and q2 indi- cate x and θ. the Lagrangian equation with respective to the generalized coordinate qi [3] can be obtained as   d @La @La  ¼ Ti . 2. and Ti is the external force. La = K − P (K means the system kinetic energy and P denotes the system potential energy. • The payload is regarded as a material particle. 2. v is a vector and it denotes the payload velocity. the system kinetic energy in Fig. • Compared with the payload mass.). 2. the rope mass is ignored. According to the assumption that the payload is regarded as a material particle. ð2:1Þ dt @ q_ i @qi where i = 1. Using Lagrangian method. qi is the generalized coordination (here.1 Structure of the single-pendulum-type overhead crane system a pendulum-type motion. Note that the payload is assumed to be a particle such that its moment of inertia is not considered in (2. When it is failed .1 can be depicted as 1 1 K ¼ M x_ 2 þ mv2 ð2:2Þ 2 2 here.52 2 Crane Mathematic Model Fig. the following assumptions [2] are given. defined as v2 ¼ v2x þ v2y . • The payload moves on the x–y surface. • The rope is considered as an inflexible rod. ð2:3Þ where vx ¼ x_ þ Lh_ cos h and vy ¼ Lh_ sin h. • The trolley moves in the x-direction.2). respectively).

5) yields @La ¼ m½ð_x þ Lh_ cos hÞðLh_ sin hÞ @h þ ðLh_ sin hÞðLh_ cos hÞ  mgL sin h ð2:10Þ ¼ mL_xh_ sin h  mgL sin h Differentiating La with respect to h_ in (2. the system potential energy in Fig. Differentiating La with respect to θ in (2.5) yields @La ¼ M x_ þ mð_x þ Lh_ cos hÞ ð2:7Þ @ x_ Further. Owing to this fact.1. 2. Differentiating La with respect to x in (2. La has the form 1 1 La ¼ K  P ¼ M x_ 2 þ mv2  mgLð1  cos hÞ ð2:5Þ 2 2 Consider the variable x.2) and (2. g is the gravitational acceleration.1 is only exhibited by the potential energy of the payload subsystem.7) with respect to time t can have    ::  d @La :: ¼ M x þm x þL€h cos h  Lh_ 2 sin h ð2:8Þ dt @ x_ Finally. its moment of inertia has to be taken into considerations. the potential energy of the trolley subsystem is kept unchanged.5) yields @La ¼m½ð_x þ Lh_ cos hÞðL cos hÞ þ ðLh_ sin hÞðLh_ sin hÞ @ h_ ð2:11Þ ¼mL_x cos h þ mL2 h_ . 2. defined as P ¼ mgLð1  cos hÞ ð2:4Þ Here.5) yields @La ¼0 ð2:6Þ @x Differentiating La with respect to x_ in (2.4). From Fig. differentiating (2.2. From (2.1 Modeling of Single-Pendulum-Type Cranes 53 to assume that the payload is a particle. the Lagrangian equation with respective to x has the form     d @La @La ::  ¼ ðm þ M Þ x þmL €h cos h  h_ 2 sin h ¼ f ð2:9Þ dt @ x_ @x Consider the variable θ.

and fi and bi (i = 1. x1 ¼ x. x2 . x4 is the angular velocity of the load.9) and (2. differentiating (2.54 2 Crane Mathematic Model Further.14) and (2. In (2. x4 T . .16). the dynamic model [4] of this overhead crane system with respect to x and θ can be obtained by means of the Lagrangian method. u is the control input.11) with respect to time t can have   d @La :: ¼ mL x cos h  mL_xh_ sin h þ mL2 €h ð2:12Þ dt @ h_ Finally. As far as state-variable-based control methods are concerned. formulated as x_ 1 ¼ x2 x_ 2 ¼ f1 ðxÞ þ b1 ðxÞu ð2:16Þ x_ 3 ¼ x4 x_ 4 ¼ f2 ðxÞ þ b2 ðxÞu Here. x2 is the trolley velocity. 2) are described as MLx24 sin x3 þ mg sin x3 cos x3 f1 ðxÞ ¼ M þ m sin2 x3 1 b1 ðxÞ ¼ M þ m sin2 x3 ðM þ mÞg sin x3 þ mLx24 sin x3 cos x4 f2 ðxÞ ¼ ðM þ m sin2 x3 ÞL cos x3 b2 ðxÞ ¼ ðM þ m sin2 x3 ÞL Equation (2. four state variables can depict this dynamic system. x3 ¼ h. the four states can be employed and a diversity of control approached can be achieved. x3 . :: ðm þ MÞ x þmLð€h cos h  h_ 2 sin hÞ ¼ f ð2:14Þ :: x cos h þ L€h þ g sin h ¼ 0 ð2:15Þ Further.13).15) can be transformed to the following state space model [5]. the Lagrangian equation with respective to θ has the form   d @La @La ::  ¼ mL x cos h þ mL2 €h þ mgL sin h ¼ 0 ð2:13Þ _ dt @ h @h From (2. x ¼ ½x1 .16) formulates the state space model of this single-pendulum-type overhead crane system. the above dynamic model composed of (2.

19) enter the dynamic model (2. it is impossible to avoid uncertainties and external disturbance in real dynamical systems. the dynamic model of the overhead crane in Fig. In the case that there are unmatched uncertainties.2 Model with Uncertainties Uncertainties can be categorized as matched uncertainties and unmatched uncer- tainties [6].18) can be written as Dfi0 ðxÞ ¼ bi ðxÞD~fi ðxÞ Dbi0 ðxÞ ¼ bi ðxÞD~bi ðxÞ Substituting (2. respectively.17) yields   x_ 2 ¼ f1 ðxÞ þ b1 ðxÞ u þ D~b1 ðxÞu þ D~f1 ðxÞ   ð2:19Þ x_ 4 ¼ f2 ðxÞ þ b2 ðxÞ u þ D~b2 ðxÞu þ D~f2 ðxÞ Apparently. etc. 2.18) into (2. overhead crane systems often are operated under uncertainty conditions such as parameter variations. the terms depicting modeling errors and parameter variations.17). Considering the possible effects of these uncer- tainties. Dfi0 ðxÞ and Dbi0 ðxÞ are assumed to be differentiable with respect to time t. Due to imperfect modeling and effects of environment. where fi(x) and bi(x) are the nominal parts of fi0 ðxÞ and bi0 ðxÞ.16). In (2. bi0 ðxÞ ¼ bi ðxÞ þ Dbi ðxÞði ¼ 1.16) is ideal and it contains no uncertainties. unmodeled dynamics. Such an entering tunnel makes this kind of uncertainties resistible by suitable control methods.17) fi0 ðxÞ ¼ fi ðxÞ þ Dfi ðxÞ. skidding and slipping. the so-called matched uncertainties mean Dfi0 ðxÞ and Dbi0 ðxÞ2spanfbi ðxÞg ð2:18Þ Explicitly. (2.2. 2Þ. Both fi(x) and bi(x) are formulated in (2. Without loss of generality. it is challenging to suppress them because it is hard to formulate these kinds of uncertainties. In reality. The uncertainties are matched if and only if the uncertainties enter a dynamical system from the control tunnel. all the uncertainties in (2. A common .1.1 Modeling of Single-Pendulum-Type Cranes 55 Note that the model (2.17) by the control tunnel. 2. indicating that they are matched.1 can have the form x_ 1 ¼ x2 x_ 2 ¼ f10 ðxÞ þ b10 ðxÞu ð2:17Þ x_ 3 ¼ x4 x_ 4 ¼ f20 ðxÞ þ b20 ðxÞu In (2.

23) can be formulated as ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi r m g xn ¼ 1þ ð2:24Þ M L . uÞ Here. uÞ ¼¼ Df2 ðxÞ þ Db2 ðxÞu. x_ 1 ¼ x2 x_ 2 ¼ f1 ðxÞ þ b1 ðxÞu þ n1 ðx. uÞ ð2:20Þ x_ 3 ¼ x4 x_ 4 ¼ f2 ðxÞ þ b2 ðxÞu þ n2 ðx. (2.23) is a second order ordinary differential equation.21) yields f :: þ L h þ gh ¼ 0 ð2:23Þ Mþm Equation (2. Consequently. (2. Substituting :: x ¼ f =ðm þ M Þ into (2. f = (m + M) x can be obtained. Because θ = 0 is the sole stable equilibrium of the overhead crane system. the angular frequency describing the oscillation of the linearized Eq.56 2 Crane Mathematic Model approach is to apply the available controllers as if there were no unmatched uncertainties.15). The unmatched uncertainties are then required to be smaller than the threshold value so that a stability result holds locally with respect to the size of the uncertainties. Finally.23). uÞ ¼ Df1 ðxÞ þ Db1 ðxÞu and n2 ðx.3 Linearized Model The nonlinear single-pendulum-type overhead crane model has been discussed and it is composed of (2. 2. This method will inevitably result in a threshold on the size of the unmatched uncertainties. n1 ðx.1.14) and (2. the following dynamic model [7] with unmatched uncertainties can be directly formulated by (2. it is important to suppress them to guarantee the system stability in the presence of significant unmatched uncertainties. The Laplace transform can be employed to solve Eq. Since unmatched uncertainties are common in control practice.20) without further simplification. The linearized equations can be written as :: :: ðm þ MÞ x þ mL h ¼ f ð2:21Þ :: :: x þ L h þ gh ¼ 0 ð2:22Þ :: According to Newton’s second law. both of the equations can be linearized around the point.

2. and payload. 2. the frequency value has a strong dependence on the mass ratio when the rope length is short. It is of interest to investigate how the frequency changes as a function of the system parameters. x (m). cable length between trolley and hook l1 (m). hook mass m1 (kg). Other symbols in Fig. and mass-point payload.4. mass-point hook.1 Modeling Figure 2. described by trolley position with respect to the origin.24).2 Variation of the system frequency Equation (2. 2. The MATLAB programs of the example are given in Appendix A.. i. The following Lagrangian equation with respect to the generalized coordinate qi can be obtained as . That is. From Fig. Figure 2.2. 2.1. On the other hand. the frequency changes very little with respect to the mass ratio when the rope length is more than 4 m. trolley. Such information can be used for physical insights of the overhead crane system. Consider the following ideal assumptions like no friction. there exist three variables to describe the crane system. The crane is moved by a driven force F. applied to the trolley. massless cables.2 demonstrates the function revealed in (2.2. To obtain the dynamic model of this crane system.4 Modeling of Double-Pendulum-Type Cranes 2.1 Modeling of Single-Pendulum-Type Cranes 57 Fig. and cable length between hook and payload l2 (m). payload mass m2 (kg).3 illustrates the schematic representation of a double-pendulum-type crane. This system consists of three subsystems.24) reveals the system frequency depending on the rope length and the mass ratio. and payload angle with respect to the vertical line θ2 (rad).1. hook. hook angle with respect to the vertical line θ1 (rad).3 are explained as trolley mass m0 (kg). Each subsystem possesses one variable. the Lagrangian method is also adopted.e.

2. the vectors v1 and v2 denote the hook and payload velocities.). the potential energy of the trolley subsystem is kept unchanged. respectively. q2. 2. 3. the system kinetic energy in Fig. defined as Pd ¼ m1 gl1 ð1  cos h1 Þ þ m2 g½l1 ð1  cos h1 Þ þ l2 ð1  cos h2 Þ ð2:28Þ Here. From Fig. 2.25).3 can be written as 1 1 1 Kd ¼ m0 x_ 2 þ m1 v21 þ m2 v22 ð2:26Þ 2 2 2 here. and q3 indicates x. They are defined as v21 ¼ v2x1 þ v2y1 ð2:27Þ v22 ¼ v2x2 þ v2y2 . qi is the generalized coordination (here.3. θ1. 2.3 Schematic of the double-pendulum-type overhead crane system   d @Lad @Lad  ¼ Ti ð2:25Þ dt @ q_ i @qi In (2. vy1 ¼ l1 h_ 1 sin h1 .58 2 Crane Mathematic Model Fig. and Ti is the external force. i = 1. the system potential energy in Fig. Lad has the form . g is the gravitational acceleration. where vx1 ¼ x_ þ l1 h_ 1 cos h1 . 2. q1. Then. and θ2. vx2 ¼ x_ þ l1 h_ 1 cos h1 þ l2 h_ 2 cos h2 and vy2 ¼ l1 h_ 1 sin h1  l2 h_ 2 sin h2 . respectively). Lad = Kd − Pd (Kd means the system kinetic energy and Pd denotes the system potential energy.3 is only exhibited by the potential energies of the hook and payload subsystems. According to the aforementioned assumptions that the payload and hook are regarded as mass-points. Owing to this fact.

1 Modeling of Single-Pendulum-Type Cranes 59 Lad ¼ Kd  Pd 1 1 1 ¼ m0 x_ 2 þ m1 v21 þ m2 v22 ð2:29Þ 2 2 2  m1 gl1 ð1  cos h1 Þ  m2 g½l1 ð1  cos h1 Þ þ l2 ð1  cos h2 Þ Consider the variable x. the Lagrangian equation with respective to x has the form   d @Lad @Lad  ¼ ðm0 þ m1 þ m2 Þ€x þ ðm1 þ m2 Þl1 €h1 cos h1 dt @ x_ @x  ðm1 þ m2 Þl1 h_ 2 sin h1 þ m2 l2 h€2 cos h2 1 ð2:33Þ  m2 l2 h_ 22 sin h2 =F Consider the variable θ1. Differentiating Lad with respect to θ1 in (2.29) yields @Lad ¼ ðm0 þ m1 þ m2 Þ_x þ ðm1 þ m2 Þl1 h_ 1 cos h1 þ m2 l2 h_ 2 cos h2 ð2:31Þ @ x_ Further.31) with respect to time t can have   d @Lad @Lad  ¼ ðm0 þ m1 þ m2 Þ€x þ ðm1 þ m2 Þl1 €h1 cos h1 dt @ x_ @x ð2:32Þ  ðm1 þ m2 Þl1 h_ 2 sin h1 þ m2 l2 h€2 cos h2 1  m2 l2 h_ 22 sin h2 Finally.29) yields @Lad ¼0 ð2:30Þ @x Differentiating Lad with respect to x_ in (2.29) yields @Lad ¼ ðm1 þ m2 Þl1 x_ h_ 1 sin h1 @h1 ð2:34Þ  m2 l1 l2 h_ 1 h_ 2 sinðh1  h2 Þ  ðm1 þ m2 Þgl1 sin h1 Differentiating Lad with respect to h_ 1 in (2.2. Differentiating Lad with respect to x in (2.29) yields @Lad ¼ ðm1 þ m2 Þðl1 x_ cos h1 þ l21 h_ 1 Þ þ m2 l1 l2 h_ 2 cosðh1  h2 Þ ð2:35Þ _ @ h1 . differentiating (2.

the Lagrangian equation with respective to θ1 has the form d @Lad @Lad ð Þ ¼ ðm1 þ m2 Þl1€x cos h1 þ ðm1 þ m2 Þl21 €h1 dt @ h_ 1 @h1 þ m2 l1 l2 h€2 cosðh1  h2 Þ þ m2 l1 l2 h_ 2 sinðh1  h2 Þ 2 ð2:37Þ þ ðm1 þ m2 Þgl1 sin h1 =0 Consider the variable θ2.29) yields @Lad ¼ m2 l1 x_ h_ 2 sin h2 þ m2 l1 l2 h_ 1 h_ 2 sinðh1  h2 Þ  m2 gl2 sin h2 ð2:38Þ @h2 Differentiating Lad with respect to h_ 2 in (2. the Lagrangian equation with respective to θ2 has the form   d @Lad @Lad  ¼ m2 l2€x cos h2 þ m2 l22 €h2 þ m2 l1 l2 €h1 cosðh1  h2 Þ dt @ h_ 2 @h2 ð2:41Þ  m2 l1 l2 h_ 21 sinðh1  h2 Þ þ m2 gl2 sin h2 =0 .60 2 Crane Mathematic Model Further. Differentiating Lad with respect to θ2 in (2.29) yields @Lad ¼ m2 l1 x_ cos h1 þ m2 l22 h_ 2 þ m2 l1 l2 h_ 1 cosðh1  h2 Þ ð2:39Þ @ h_ 2 Further.39) with respect to time t can have   d @L ¼ m2 l2€x cos h2  m2 l2 x_ h_ 2 cos h2 þ m2 l22 €h2 þ m2 l1 l2 €h1 cosðh1  h2 Þ dt @ h_ 2  m2 l1 l2 h_ 21 sinðh1  h2 Þ þ m2 l1 l2 h_ 1 h_ 2 sinðh1  h2 Þ ð2:40Þ Finally. differentiating (2. differentiating (2.35) with respect to time t can have    ::  d @Lad ¼ ðm1 þ m2 Þ l1 x cos h1  l1 x_ h_ 1 cos h1 þ l21 €h1 dt @ h_ 1 ð2:36Þ þ m2 l1 l2 €h2 cosðh1  h2 Þ  m2 l1 l2 h_ 1 h_ 2 sinðh1  h2 Þ þ m2 l1 l2 h_ 2 sinðh1  h2 Þ 2 Finally.

q_ Þ ¼ 4 0 0 m2 l1 l2 h_ 2 sinðh1  h2 Þ 5 0 m2 l1 l2 h_ 1 sinðh1  h2 Þ 0 GðqÞ ¼ ½ 0 ðm1 þ m2 Þgl1 sin h1 m2 gl2 sin h2 T Further. g is the gravitational acceleration. and (2. Cðq. (2. and θ2 [8] can be obtained by means of the Lagrangian method.41). M _ q_ is a vector of Coriolis and centripetal torques. ðm0 þ m1 þ m2 Þ€x þ ðm1 þ m2 Þl1 €h1 cos h1 þ m2 l2 €h2 cos h2 ð2:42Þ  ðm1 þ m2 Þl1 h_ 21 sin h1  m2 l2 h_ 22 sin h2 ¼ F ðm1 þ m2 Þl1€x cos h1 þ ðm1 þ m2 Þl21 €h1 þ m2 l1 l2 €h2 cosðh1  h2 Þ ð2:43Þ þ m2 l1 l2 h_ 2 sinðh1  h2 Þ þ ðm1 þ m2 Þgl1 sin h1 ¼ 0 2 m2 l2€x cos h2 þ m2 l1 l2 €h1 cosðh1  h2 Þ þ m2 l22 €h2 ð2:44Þ  m2 l1 l2 h_ 2 sinðh1  h2 Þ þ m2 gl2 sin h2 = 0 1 Rearrange (2. (2.37). The expression [9. qÞ and G(q) is a vector of the gravitational term. Cðq.42). θ1. and (2.33). h1 .43). 0T is a vector of the generalized force. (2.1 Modeling of Single-Pendulum-Type Cranes 61 From (2.2. qÞ _ and G(q) are defined as 2 3 m0 þ m1 þ m2 ðm1 þ m2 Þl1 cos h1 m2 l2 cos h2 MðqÞ ¼ 4 ðm1 þ m2 Þl1 cos h1 ðm1 þ m2 Þl21 m2 l1 l2 cosðh1  h2 Þ 5 m2 l2 cos h2 m2 l1 l2 cosðh1  h2 Þ m2 l22 2 3 0 ðm1 þ m2 Þl1 h_ 1 sin h1 m2 l2 h_ 2 sin h2 Cðq. h2 T is a vector of the three generalized coordinates. the dynamic model of the double-pendulum-type overhead crane system with respect to x. The three equations can be rewritten as :: MðqÞ q þ Cðq.44) in the form of a matrix. 10] has the form . q_ Þq_ þ GðqÞ ¼ s ð2:45Þ Here. s ¼ ½F. MðqÞ. q ¼ ½x.45) can be transformed to its state space expression. (q) is a 3 × 3 inertia matrix. 0.

Here Γi. and fi(x) and bi(x) (i = 1. x4. x5. x2 is the trolley velocity. x3 = θ1.62 2 Crane Mathematic Model x_ 1 ¼ x2 x_ 2 ¼ f1 ðxÞ þ b1 ðxÞu x_ 3 ¼ x4 ð2:46Þ x_ 4 ¼ f2 ðxÞ þ b2 ðxÞu x_ 5 ¼ x6 x_ 6 ¼ f3 ðxÞ þ b3 ðxÞu In (2. x6 is the angular velocity of the payload. u = F is the control input. x2. formulated by fi (x) = Γi/Δ and bi(x) = Τi/Δ. x1 = x. x3. 2. x4 is the angular velocity of the hook. 3) are nonlinear functions of the vector x. x6]T.46). and Δ are determined by   D = ðm1 þ m2 Þm2 l21 l22 ðm0 þ m1 þ m2 Þ  ðm1 þ m2 Þ cos2 ðx3 Þ   m22 l21 l22 ðm1 þ m2 Þ cos2 ðx5 Þ þ ðm0 þ m1 þ m2 Þ cos2 ðx3  x5 Þ 2ðm1 þ m2 Þ cosðx3 Þ cosðx5 Þ cosðx3  x5 Þ   C1 ¼ ðm1 þ m2 Þm2 l21 l22  m22 l21 l22 cos2 ðx3  x5 Þ ðm1 þ m2 Þl1 x24 sinðx3 Þ   þ m2 l2 x26 sinðx5 Þ þ ðm1 þ m2 Þm2 l1 l22 cosðx3 Þ   m22 l1 l22 cosðx5 Þ cosðx3  x5 Þ m2 l1 l2 x26 sinðx3  x5 Þ  þðm1 þ m2 Þgl1 sinðx3 Þ þ ðm1 þ m2 Þm2 l21 l2 cosðx5 Þ   m2 l21 l2 cosðx3 Þ cosðx3  x5 Þ m2 l1 l2 x24 sinðx3  x5 Þ þ m2 gl2 sinðx5 Þ T1 ¼ ðm1 þ m2 Þm2 l21 l22  m22 l21 l22 cos2 ðx3  x5 Þ   C2 ¼ m22 l1 l22 cosðx5 Þ cosðx3  x5 Þ  ðm1 þ m2 Þm2 l1 l22 cosðx3 Þ ðm1 þ m2 Þl1 x24 sinðx3 Þ    þm2 l2 x26 sinðx5 Þ þ m22 l22 cos2 ðx5 Þ  ðm0 þ m1 þ m2 Þm2 l22 m2 l1 l2 x26 sinðx3  x5 Þ þðm1 þ m2 Þgl1 sinðx3 Þ þ ½ðm0 þ m1 þ m2 Þm2 l1 l2 cosðx3  x5 Þ   ðm1 þ m2 Þm2 l1 l2 cosðx3 Þ cosðx5 Þ m2 l1 l2 x24 sinðx3  x5 Þ þ m2 gl2 sinðx5 Þ T2 ¼ m22 l21 l22 cosðx5 Þ cosðx3  x5 Þ  ðm1 þ m2 Þm2 l1 l22 cosðx3 Þ C3 ¼ ðm1 þ m2 Þm2 l21 l2 ½cosðx3 Þ cosðx3  x5 Þ  cosðx5 Þ ½ðm1 þ m2 Þl1 x24 sinðx3 Þ  þm2 l2 x26 sinðx5 Þ þ ½ðm0 þ m1 þ m2 Þm2 l1 l2 cosðx3  x5 Þ   ðm1 þ m2 Þm2 l1 l2 cosðx3 Þ cosðx5 Þ m2 l1 l2 x26 sinðx3  x5 Þ þ ðm1 þ m2 Þgl1 sinðx3 Þ h i þ ðm1 þ m2 Þ2 l21 cos2 ðx3 Þ  ðm0 þ m1 þ m2 Þðm1 þ m2 Þl21    m2 l1 l2 x24 sinðx3  x5 Þ þ m2 gl2 sinðx5 Þ . Τi. the vector x is defined by [x1. x5 = θ2.

The analysis is very similar to the process in Sect. The linearized crane model can be written as . It can be treated as the nominal model of double-pendulum-type overhead cranes. Usually. 2.46). the moment of inertia of the payload subsystem cannot be ignored and it has to be taken into consideration.13) can be linearized around θ1 = 0 and θ2 = 0. uÞ x_ 5 ¼ x6 x_ 6 ¼ f3 ðxÞ þ b3 ðxÞu þ n3 ðx. the mass-point assumption for the payload subsystem can be satisfied under some operating conditions. 2. uÞ where i = 1. if it can be written as ni ðx. uÞ in (2.1.2.13). the three uncertain terms are unmatched because they cannot enter the crane model by the control channel. uÞ x_ 3 ¼ x4 ð2:47Þ x_ 4 ¼ f2 ðxÞ þ b2 ðxÞu þ n2 ðx. Note that the three uncertain terms have to be treated as a whole.47) is matched. Considering the system uncertainties. Otherwise.5 Model with Uncertainties Equation (2.1. uÞ The uncertain part ni ðx. uÞ ¼ bi ðxÞDni ðx. the uncertain equations can be derived from (2.2. if only a part of the three terms can enter the crane model by the control channel. Because θ1 = θ2 = 0 is the sole stable equilibrium of the double-pendulum-type crane system. However. Concerning these extreme operating conditions. Briefly. The uncertainties are still unmatched.1. 2. (2. 3. 2.1 Modeling of Single-Pendulum-Type Cranes 63 T3 ¼ ðm1 þ m2 Þm2 l21 l2 cosðx3 Þ cosðx3  x5 Þ  ðm1 þ m2 Þm2 l21 l2 cosðx5 Þ Note that the model (2. 2.45) has two important assumptions that are mass-point hook and payload. the mass-point assumption can be satisfied for the hook subsystem. the uncertain model of the double-pendulum-type overhead crane in Fig.46) is an ideal model.1 can be described as x_ 1 ¼ x2 x_ 2 ¼ f1 ðxÞ þ b1 ðxÞu þ n1 ðx.6 Linearized Model The nonlinear double-pendulum-type overhead crane model is shown in (2.

It corresponds closely to the frequency of a single pendulum with the length of l1 + l2.48). Their expressions [11] are determined as rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi g pffiffiffi x1 ¼ a b ð2:49Þ 2 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi g pffiffiffi x2 ¼ aþ b ð2:50Þ 2 Here.4.49) and (2. α has a form   m1 þ m2 1 1 a¼ þ m1 l1 l2 Another parameter β is formulated as  2  2   m1 þ m2 1 1 m1 þ m2 1 b¼ þ 4 m1 l1 l2 m1 l1 l2 From (2. the two natural frequencies not only depend on the length of the cables but also depend on the masses of payload and hook. the cable length between hook and payload l2 is considered as a variable when the total length l determined by l1 plus l2 is held constant at 6 m. R = m2/m1 is defined as the payload-to-hook mass ratio. On the other hand.50). 2. the matrixes M 2 3 m0 þ m1 þ m2 ðm1 þ m2 Þl1 m2 l2  ¼ 4 ðm1 þ m2 Þl1 M ðm1 þ m2 Þl21 m2 l1 l2 5 m2 l2 m2 l1 l2 m2 l22 2 3 0 0 0 K ¼ 4 0 ðm1 þ m2 Þgl1 0 5 0 0 m2 gl2 The two natural frequencies of the double-pendulum-type overhead crane system can be obtained by the nonzero eigenvalues of the matrix—M K. To simplify this problem. Figure 2. but it can be dramatically changed by the hoisting operation [12]. In Fig. It is interesting to investigate how the frequencies change as a function of the system physical parameters. 2 . ω1 changes very little for a constant l1 + l2.4 illustrates the two oscillation frequencies as a function of R and l2.64 2 Crane Mathematic Model  q:: þKq ¼ 0 M ð2:48Þ  and K are determined as In the linearized model (2. the value of ω1 is maximized for a constant l1 + l2 when the two cable lengths are equal to l1 þl 2 .

01 to 2.4 Variation of the two frequencies Furthermore.1:0. the value of ω2 varies substantially more than the value of ω1. % K and L arrays for 3-D plots. Appendices A Matlab Codes to Plot Fig.w) % 3-D colored surface.01:0. low payload-to-hook mass ratios and equal cable lengths are more representative to depict the double-pendulum motions of the crane for the point-to-point transport control with a constant l1 + l2.L.L]=meshgrid(k./L). 2.1 Modeling of Single-Pendulum-Type Cranes 65 Fig.1 to 10.2 k=0. the high-frequency -2 has a strong dependence on the cable length l2.8.01:2.01. % Mass ratio from 0. In brief. but the high-frequency ω2 becomes more important to the double-pendulum motions for low payload-to-hook mass ratios. every each 0.l).1 [K. w=sqrt((1+K)*9. % Rope length from 0. l=0.2. R has a relatively small effect on ω1. every each 0. the contribution of ω2 to problematic swing amplitude is particularly large for a constant l1 + l2 when the two cable lengths are approximately equal.1:10. . Concerning the mass ratio. % Array of angular frequency surf(K. 2.

2907363 . Oguamanam D. Spong MW.^2.L2. % R and L2 arrays for 3-D plots p=sqrt((1+R)./L2)+p).1:6.1:0. Control Eng Pract 4(5):645–653 7. Guo W. Lee SG (2013) Sliding mode controls of double-pendulum crane systems. Kim D.L2]=meshgrid(r.*((1+R). Zhao D. Cheng C. New York 4. Chen CY (1996) Controller design for an overhead crane system with uncretainty. Liu D.1:0. Wang W (2005) Adaptive sliding mode fuzzy control for a two-dimensional overhead crane.8.w1) %3-D colored surface hold.*(1. Lee HH (1998) Modeling and control of a three-dimensional overhead crane. J Sound Vib 242(3):411–426 5. %cable length between hook and payload l1=6-l2.66 2 Crane Mathematic Model B Matlab Codes to Plot Fig. Mech Sci 4:251–261 11. Chwa D. Vaughan J./(L2. J Dyn Syst Meas Control Trans ASME 130(3): doi:10. Heppler G (2001) Dynamics of a three-dimensional overhead crane system. %mass ratio l2=0./(6-L2)+1./(6-L2)+1. Kenison M (2008) Input shaping control of double-pendulum bridge crane oscillations. Nayfeh AH.*sqrt((1+R). % gravitational acceleration [R. w1=sqrt(g/2). Wiley.*sqrt((1+R).*(1.4 r=0./(6-L2)+1. IEEE Trans Ind Electron 55(11):3972–3984 8. Hong SK (2008) Antisway tracking control of overhead cranes with system uncertainty and actuator nonlinearity using an adaptive fuzzy sliding-mode control. O’Connor W. Abdel-Rahman EM. Hansen JS. Park MS. Singhose W. distributed-mass load using mechanical wave concepts.*(1. J Vib Control 9(7):863–908 2. %l1 plus l2 is held constant at 6m g=9.L2. J Dyn Syst Meas Control Trans ASME 120(4):471–476 3.*(6-L2))))./L2)-p). % Another surface surf(R. Habibi H (2013) Gantry crane control of a double-pendulum. Kim D. Mechatronics 15(5):505–522 6. J Mech Sci Technol 27(6):1863–1873 9. Singhose W (2010) Control of tower cranes with double-pendulum payload dynamics. Vidyasagar M (2006) Robot modeling and control. Liu D. Hutchinson S. Yi J.^2-4. Yi J (2008) Dynamics and GA-based stable control for a class of underactuated mechanical systems. Int J Control Autom Syst 6(1):35–43 10. Tuan LA.w2) %3-D colored surface References 1./L2). IEEE Trans Control Syst Technol 18(6):1345–1358 12. 2. % Array of w2 surf(R. % Array of w1 w2=sqrt(g/2).l2).1:2. Masoud ZN (2003) Dynamics and control of cranes: a review.1115/1.

In addition. the chapter proceeds with source codes of all the simulations. integral SMC. payload swing reduction. the two aspects contradict each other. Meanwhile. Inherently. For the purpose of illustration. DOI 10.e. To focus on the SMC methods. Keywords Overhead crane control  First-order SMC  Integral SMC  Terminal  SMC Second-order SMC 3. i. The control task has two aspects. a control mechanism or algorithm has to be applied to crane systems. and crane scheduling. Concerning transport control of a crane system. trolley posi- tioning. its control task [2] is to transport the payload to a desired position as accurately and quickly as possible.Chapter 3 Overhead Crane Control by Sliding Mode Methods Abstract The methodology of sliding mode control (SMC) covers a diversity of design methods. and second-order SMC. i. The physical properties of cranes cannot simultaneously realize the two aspects. where the crane scheduling problem should resort to optimization of crane operations and the other two control problems can be dealt with control algorithms. i.. fast crane posi- tioning and small payload swing. Moreover. To solve the inherent contradiction and achieve the transport task of cranes. This chapter introduces several typical SMC design methods. its control system should suppress payload oscillations no matter what kind of pendulum-type they are. the nominal models of both types of overhead cranes are considered during control design.1 Problem Description Overhead.e.1007/978-3-662-48417-3_3 .e.. crane control includes several control problems [1]. the uncertain models are also taken to analyze the system stability. both single-pendulum-type and double-pendulum-type overhead cranes are adopted as research benchmarks and these SMC methods are carried out by the transport control problem of cranes. Qian and J. © Springer-Verlag Berlin Heidelberg 2016 67 D. Yi. terminal SMC.. Hierarchical Sliding Mode Control for Under-actuated Cranes. first-order SMC.

3. the task of the transport control problem fuses both the trolley positioning and the payload swing reduction. the design of the equivalent-control-based SMC is adopted in the rest of the book. overhead crane control just means the transport control problem without independent claim. where the SMC design based on state feedback is adopted in Chap.2 is applied to two overhead cranes. the first-order SMC method is a basic one. x in (2.16) is the state vector.68 3 Overhead Crane Control by Sliding Mode Methods As mentioned. and c is the parameter vector of this sliding surface and it should be predefined. As a result. i. Recall the equivalent-control-based SMC law (1. Design the control law of the crane system (1. Provided that all the crane states are measurable.16) in Chap. Since the sliding mode control (SMC) methodology is a feedback control design. the state feedback-based SMC design is adopted in the rest of the book without independent claim. the transport control problem of crane systems is representative.16) into the derivative of s yield s_ ¼ cT x_ ¼ c1 x_ 1 þ c2 x_ 2 þ c3 x_ 3 þ c4 x_ 4 ¼ c1 x2 þ c2 ½f1 ðxÞ þ b1 ðxÞu þ c3 x4 þ c4 ½f2 ðxÞ þ b2 ðxÞu ð3:2Þ ¼ c1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ þ c2 b1 ðxÞu þ c4 b2 ðxÞu . output feedback design and state feedback design. but it is important and representative.20) in Chap.1 Control Design of Single-Pendulum-Type Overhead Cranes Recall the model of single-pendulum-type overhead crane systems (2.2. 3. there are two ways to design a SMC system [3]. Recall Chap. 2. To illustrate its design.e. 2. 1. the first-order SMC method in Sect.20). In this chapter. described by the equivalent control law ueq plus the switching control law usw. Owing to its advantages.. 3. To obtain its equivalent control law. Design a linear sliding surface of such a crane as s ¼ cT x ð3:1Þ where s is the sliding surface variable.2 First-Order Sliding Mode Control Among a diversity of sliding mode design methods. they are directly adopted for control design without more explanations. differentiating s in (3. In this book.1) with respect to time t and substituting the crane model (2. The cranes cover both single-pendulum-type and double-pendulum-type crane systems. The mathematic models of the two types of cranes have been established. 2.

a Lyapunov function candidate in the Lyapunov’s stability scheme can be defined by 1 VðtÞ ¼ s2 ð3:5Þ 2 Differentiating V with respect to time t and substituting the crane model (1.2 Stability Analysis of the Single-Pendulum-Type Crane Control System Concerning the crane control system based on the first-order SMC.2 First-Order Sliding Mode Control 69 When the system states keep sliding along the sliding surface (3.4). define .6) yields _ VðtÞ ¼ susw ½c2 b1 ðxÞ þ c4 b2 ðxÞ ð3:7Þ _ From (3.3.1).20) into the derivative of V yield _ VðtÞ ¼ s_s ¼ sðc1 x_ 1 þ c2 x_ 2 þ c3 x_ 3 þ c4 x_ 4 Þ ¼ sfc1 x_ 1 þ c2 ½f1 ðxÞ þ b1 ðxÞu þ c3 x_ 3 þ c4 ½f2 ðxÞ þ b2 ðxÞug ¼ sfc1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ þ ½c2 b1 ðxÞ þ c4 b2 ðxÞug ¼ sfc1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ þ ½c2 b1 ðxÞ þ c4 b2 ðxÞðueq þ usw Þg ð3:6Þ Consider the expression of the equivalent control law (3. As a result. The switching control law usw is still kept unknown. the sole equivalent control law is applied to the crane control system and s_ ¼ 0 exists in (3. Substituting it into (3.2. the sliding surface (3.16). only the equivalent control law ueq is obtained. the equivalent control law ueq can be in the form of c1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ ueq ¼  ð3:4Þ c2 b1 ðxÞu þ c4 b2 ðxÞ Compared with (1.5). These linguistic descriptions are formulated by s_ ¼ c1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ þ c2 b1 ðxÞu þ c4 b2 ðxÞueq ¼ 0 ð3:3Þ Consequently. In the sense of Lyapunov.2). VðtÞ  0.1). 3. and the equivalent-control-based SMC law (1.20). VðtÞ\0 should exist to make the crane control system asymptotically stable. To obtain the whole control law u. the switching control law will be deduced from the system stability analysis.

Differentiating V with respect to time t and substituting the crane model with unmatched uncertainties (2. uÞg ð3:10Þ ¼ sfc1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ þ ½c2 b1 ðxÞ þ c4 b2 ðxÞu þ n1 ðx. the sliding surface (3. the first-order SMC law of the crane system with unmatched certainties can be deduced from (3. uÞjj Finally. define j s þ g sgnðsÞ usw ¼  ð3:11Þ c2 b1 ðxÞ þ c4 b2 ðxÞ In (3. the invariance property of SMC can resist matched uncertainties.20) is taken into considerations.4). both κ* and η* are positive and constant. To test the effects of unmatched uncertainties on the system stability. uÞg ¼ sf½c2 b1 ðxÞ þ c4 b2 ðxÞusw þ n1 ðx.4) and (3.5).19). where g [ sup jjn1 ðx.8): u ¼ ueq þ usw c1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ js þ g sgnðsÞ ð3:9Þ ¼  c2 b1 ðxÞu þ c4 b2 ðxÞ c2 b1 ðxÞ þ c4 b2 ðxÞ As mentioned.1). adopt the equivalent-control-based SMC law (1.5). uÞ þ n2 ðx. uÞ þ n2 ðx.20).70 3 Overhead Crane Control by Sliding Mode Methods js þ g sgnðsÞ usw ¼  ð3:8Þ c2 b1 ðxÞ þ c4 b2 ðxÞ where both κ and η are the positive constants.4) and (3.9) is carried out by the uncertain crane system (2. uÞg _ VðtÞ  0 in (3. In the sense of Lyapunov.11): .1) and the equivalent-control-based SMC law (1.20) into the derivative of V yield _ VðtÞ ¼ s_s ¼ sðc1 x_ 1 þ c2 x_ 2 þ c3 x_ 3 þ c4 x_ 4 Þ ¼ sfc1 x_ 1 þ c2 ½f1 ðxÞ þ b1 ðxÞu þ n1 ðx. the crane model with unmatched uncertainties (2. where the equivalent control law of the nominal crane model is determined by (3.19). and select the Lyapunov function candidate (3. VðtÞ\0 should exist to make the uncertain crane control system asymptotically stable. Define the sliding surface (3. uÞ þ c3 x_ 3 þ c4 ½f2 ðxÞ þ b2 ðxÞu þ n2 ðx. Finally. uÞ þ n2 ðx. As a result. the first-order SMC law of the single-pendulum-type overhead crane system can be deduced from (3.11). indicating that the system stability will be kept unchanged if the control law (3.

2d.1 and 3. the sliding motion begins at about 1 s.2 First-Order Sliding Mode Control 71 u ¼ ueq þ usw c1 x2 þ c3 x4 þ c2 f1 ðxÞ þ c4 f2 ðxÞ j s þ g sgnðsÞ ð3:12Þ ¼  c2 b1 ðxÞu þ c4 b2 ðxÞ c2 b1 ðxÞ þ c4 b2 ðxÞ Compared with (3. Thereafter.11) needs a larger coefficient of the signum function to resist the adverse effects of unmatched uncertainties on the system stability. From Fig. the crane system spends about 3 s arriving at the desired position in Fig.3. 3.2c and the switching control law in Fig. 3. the switching control law is kept zeroth in Fig. the control law for crane systems with unmatched uncertainties (3.9) is adopted.81 Initial state vector x0 [0 m 0 m s−1 0 rad 0 rad s−1] Desired state vector xd [1. 3.2d and the sole equivalent control law forces the sliding mode along the sliding surface. Physical parameters of the overhead crane system are shown in Table 3.2.1b when the control task is achieved. 3. Here.2b.2a. 3. Figures 3. the first-order SMC law is composed of the equivalent control law in Fig. 3. which is shown in Appendix A.1a. In Fig. and c = [–3 –3 10.8 Cable length L (m) 0. they are just employed for the purpose of illustration. The equivalent-control-based SMC law (3. The three curves show that the equivalent-control-based SMC method can effectively decrease the chattering. so are the initial and desired state vectors. Note that the physical parameters of the crane system in Table 3.3 Simulations of Nominal Single-Pendulum-Type Overhead Cranes To demonstrate the control performance of the first-order SMC law for nominal overhead crane systems.2. Table 3.8). they can be scaled up.2 can be obtained by a Simulink model of MATLAB.1 and 3.3 1]T after trial and error. Figure 3.1 Physical parameters and initial and desired states Trolley mass M (kg) 1 Payload mass m (kg) 0. 3. The first-order SMC law can effectively suppress the single-pendulum-type motions of the payload in Fig. where the controller parameters are determined by κ = 4.2b displays the curve of the sliding surface s.305 Acceleration of gravity g (m s−2) 9. 3. some simulation results will be displayed and some analyses will be presented. In practice.05. 3. From Figs.1 are very small compared with a real crane.0 m 0 m s−1 0 rad 0 rad s−1] . η = 0.1 [4].

3. c Payload angle. b Sliding surface s.1 System performance by the first-order SMC method for the overhead crane. 3.2 Control performance of the first-order SMC method. c Equivalent control ueq. d Payload angular velocity Fig. a Trolley position. d Switching control usw .72 3 Overhead Crane Control by Sliding Mode Methods Fig. a Control input u. b Trolley velocity.

1.2c. 1). d Payload angular velocity . uÞjj\ 0:2 and the other parameter κ* = κ = 4. 3. the payload has some slightly residual oscillations in the presence of uncertainties in Fig.2. Assume that both the uncertain terms in the model are determined by 0.4 Simulations of Uncertain Single-Pendulum-Type Overhead Cranes Recall the crane model with unmatched uncertainties (2. and the desired state vector are also kept unchanged from Table 3. Figure 3.4 shows the control performance though the control input and sliding surface curves.1 and 3. The Simulink model of MATLAB to plot Figs.3c.2. Compared with Fig.3 and 3.3 System performance by the first-order SMC method for the crane with unmatched uncertainties. c Payload angle.9) applied to the nominal crane model can perfectly suppress the single-pendulum oscillations of the payload in Fig. This fact indicates that some lightly residual oscillations in a real crane control system are evitable because of unmatched uncertainties. the initial state vector. uÞ þ n2 ðx.19). In contrast. 3. Compared with Fig. For the purpose of comparison. where rand() is a MATLAB command to generate a uniformly distributed pseudorandom number on the open interval (0. the control law (3. 3.3. the parameter vector of the sliding surface c is kept unchanged. The physical parameters of the crane. η* is set to 0. the trolley in Fig.1a. b Trolley velocity.3a can likewise achieve its positioning at about 3 s. To make the crane control system asymptotically stable.2 First-Order Sliding Mode Control 73 3. Figure 3.25 because sup jjn1 ðx. 3. 3. However.4 is very similar to that of plotting Figs. a Trolley position. 3. 3.1 × [1–2 × rand()].3 shows the crane performance though the state variable curves of the uncertain crane system.

2. indicating that the system trajectory deviates from the sliding surface due to the adverse effects of uncer- tainties. describing the nominal double-pendulum-type overhead crane system in Fig. c Equivalent control ueq.2. the parameter tuning. On basis of the model.74 3 Overhead Crane Control by Sliding Mode Methods Fig. 3.2. 3.2.2d. 3.1 and 3. b Sliding surface s. the switching control works all the time to resist the unmatched uncer- tainties in the system dynamic process because of the existence of unmatched uncertainties.3. some small jumps exist. Although the theoretical design and analysis are similar with single-pendulum-type overhead crane systems. Consider a double-pendulum-type overhead crane system as well.46). . especially the tuning about the sliding surface parameters. a Control input u. d Switching control usw Fig. Such jumps also make the control input in Fig. The control design and analysis for the double-pendulum-type crane system is very similar to Sects. In Fig. 3. Recall the dynamic model (2. adopt the standard equivalent-control-based SMC method presented in Chap. 3. 3.4b.4a jump back and forth. is a time-consuming business.5 Extensions of Double-Pendulum-Type Overhead Cranes The first-order SMC method can also be extended to double-pendulum-type overhead crane systems. 2.4 Control performance of the first-order SMC method for the crane with unmatched uncertainties. 2.

The method presented by Tuan and Lee [5] can be treated as mandatory decoupling. To obtain its equivalent control law. e = x1 – x1d and the derivative of e is equal to x2.14). their design. For the purpose of illustration.13) are ignored. their idea can be treated as an order-reduction sliding mode. The novel sliding surface is defined by s ¼ e_ þ ke þ ax3 þ bx5 ð3:13Þ In (3.13). analysis. and simulation are investigated step by step. two variables x4 and x6 in (3. α.14). β. to keep these system states on the sliding manifold. Compared with the standard equivalent-control-based SMC law. Tuan and Lee [5] designed a novel sliding surface for such a double-pendulum-type overhead crane.15) can guarantee all state trajectories slide on the sliding surface (3. This fact means that the sliding surface does not need the angular velocities of the payload and hook. Compared with the standard sliding surface.46) into the derivative of s yield s_ ¼ x_ 2 þ kx2 þ ax4 þ bx6 ð3:14Þ ¼ f1 ðxÞ þ b1 ðxÞu þ ax4 þ bx6 When the system states sliding along the sliding surface (3. the switching control law should be com- plemented such that the final control law can be written by u ¼ ðm1 þ m2 Þl1 cos x3 x_ 4  m2 l2 cos x5 x_ 6 þ ðm1 þ m2 Þl1 sin x3 x24 þ m2 l2 sin x5 x26 ð3:16Þ  ðm0 þ m1 þ m2 Þðkx2 þ ax4 þ bx6 Þ  K sgnðsÞ The standard sliding surface has six state variables such that the derivative of s covers the hook and payload controls are included in the final control law.46). s_ ¼ 0 exists and the sole equivalent control law is applied to the crane control system. these linguistic descriptions are formulated by ueq ¼ ðm1 þ m2 Þl1 cos x3 x_ 4  m2 l2 cos x5 x_ 6 þ ðm1 þ m2 Þl1 sin x3 x24 þ m2 l2 sin x5 x26 ð3:15Þ  ðm0 þ m1 þ m2 Þðkx2 þ ax4 þ bx6 Þ The equivalent control (3. the system stability has .13) when the sliding mode is reached. and λ are the sliding surface parameters.13) other than five parameters in the standard sliding surface.3.13). Consequently. differentiating s in (3. a novel design is introduced.13) with respect to time t and substituting the crane model (2. However. From (3.2 First-Order Sliding Mode Control 75 To exhibit the development of SMC. Consider the crane dynamics (2. where the sliding mode exists in a subspace of the standard sliding mode. But there is only the trolley control in (3. Such a definition can simplify the parameter tuning because only three parameters exist in (3.

There are four controller parameters in (3.42) and (2. It is interesting to investigate the inherent relationship among the three parameters of the sliding surface. define the Lyapunov function candidate as 1 V ¼ s2 ð3:17Þ 2 Differentiating V with respect to time t and substituting the Eq. One parameter K is about the switching control law and K [ 0 has been drawn from the system stability. From (3. and (3.21).15) into the derivative of V yield K V_ ¼ s_s ¼ s sgnðsÞ ð3:18Þ ðm0 þ m1 þ m2 Þ For any K [ 0. the system stability can be touched by the linearization technique. the system dynamics are composed of (3.17). The other three parameters are about the sliding surface.42) and the equivalent control law (3. (3. (2.14) such that x_ 2 ¼ kx2  ax4  bx6 ð3:19Þ Substitute (3. and (3. Apparently.19) into (2. Apparently.16). According to the Lyapunov’s indirect method. This purpose can be achieved by the dynamic analysis after the sliding mode takes place. To verify the system stability. s_ ¼ 0 exists in (3.43) yield ðm1 þ m2 Þ cos x3 ðkx2 þ ax4 þ bx6 Þ þ ðm1 þ m2 Þl1 x_ 4 þ m2 l2 cosðx3  x5 Þ_x6 þ m2 l2 sinðx3  x5 Þx26 þ ðm1 þ m2 Þg sin x3 ¼ 0 ð3:20Þ cos x5 ðkx2 þ ax4 þ bx6 Þ þ l1 cosðx3  x5 Þ_x4 ð3:21Þ þ l2 x_ 6  l1 sinðx3  x5 Þx24 þ g sin x5 ¼ 0 When the state trajectories slide on the surface and convergent to the origin. Then the linearized dynamics are written by x_ ¼ Ax ð3:22Þ . When all the state variables slide on the sliding surface. Then.21). V\0_ exists in (3.76 3 Overhead Crane Control by Sliding Mode Methods to be taken into considerations after the mandatory decoupling.18).19). rearranging (2. (3.20).18).17) and (3.43).20).19). the SMC system on basis of the novel sliding surface is of asymptotic stability in the sense of Lyapunov. Linearize (3. V  0 can be obtained in (3. all the equations are nonlinear.42) and (2.

the control law will be carried out to a double-pendulum-type overhead crane. the crane can arrive at the desired position at about 7. indicating that the trolley can directly arrive at the desired position instead of moving back and forth around the desired position. which is shown in Appendix B.2.5a has no overshooting. The controller parameters of the first-order SMC law [5] are selected by K = 70. the hook angular velocity. On the other hand.5c and e. Figure 3.2 Physical Trolley mass m0 (kg) 50 parameters of the double-pendulum-type Hook mass m1 (kg) 10 overhead crane Payload mass m2 (kg) 2 Trolley-hook cable length l1 (m) 3 Hook-payload cable length l2 (m) 0. the hook angle.5. The control performance by the first-order SMC law is shown in Fig. the driven force applied to the trolley.22) must be a Hurwitz matrix to guarantee the stability of the linearized system (3. In Fig. 3. the payload angle. Displayed in (3. the chattering phenomenon is greatly reduced because of the equivalent-control-based control design. λ = 0. 3. and the payload angular velocity are located in Fig.5a–f.3. 3. 3. Figure 3.22).18) and (3.e..6a. 3.5 s. This fact results in k[0 a[0 b\0 ð3:23Þ To illustrate the feasibility of the first-order SMC law. In addition. 3.3 Acceleration of gravity g (m s−2) 9. In Fig. obtained by a Simulink model of MATLAB. the controller can effectively resist the double-pendulum oscillations in Fig. Meanwhile.23). the control input.81 .5 and 3. Some numerical simulation results are demonstrated in Figs.5. where the crane physical parameters [5] are determined by Table 3. the signs of the four parameters coincide with the proven results. where the trolley position.5a. jumps back and forth at the outset to suppress the double-pendulum oscillations. the trolley velocity.6. From Fig. β = –11. the transport task of the double-pendulum-type crane is set as the initial state vector x0 = [0 m 0 m s−1 0 rad 0 rad s−1 0 rad 0 rad s−1] and the desired state vector xd = [4 m 0 m s−1 0 rad 0 rad s−1 0 rad 0 rad s−1]. respectively. i. 3.6b shows the curve of the novel sliding surface defined by Table 3.5 shows the curves of the crane state variables during the transport dynamics. α = 17. the curve in Fig.6. 3.2 First-Order Sliding Mode Control 77 here 2 3 0 1 0 0 0 0 6 0 k 0 a 0 b 7 6 7 60 0 0 1 0 0 7 A¼6 6 0  lk 7 6  lg1  la1 0  lb1 7 7 40 0 1 5 1 0 0 0 0  lk2 0  la2  lg2  lb2 A in (3.

e Payload angle.6b.5 System performance by the first-order SMC method for the double-pendulum-type overhead crane system. a Trolley position. b Sliding surface a part of the crane state variables. . b Trolley velocity. c Hook angle. 3.78 3 Overhead Crane Control by Sliding Mode Methods Fig. From Fig.13) and converge to the origin. d Hook angular velocity. the sliding mode is reached at about 3 s. 3. f Payload angular velocity Fig.6 Control performance by the first-order SMC method for the double-pendulum-type overhead crane system. a Control input. the state trajectories of the closed-loop crane control system by the sliding mode keep sliding along the sliding surface (3. 3. Thereafter.

To develop such an integral sliding mode controller. This characteristic makes the ISMC method attractive because the system trajectory in an ISMC system always starts from the sliding surface. driven force applied to the trolley f. x is the state vector of the single-pendulum-type overhead crane system. the integral term in the surface breaks out of this point and it promotes the dimension of the sliding mode. B. only single-pendulum-type over- head crane systems are considered. . However.e.3. and F are determined by 2 3 2 3 2 3 0 1 0 0 0 0 60 0 0 07 6 b1 ðxÞ 7 6 f1 ðxÞ 7 A¼6 40 7 B¼6 7 6 4 0 5 F¼4 0 5 7 0 0 15 0 0 0 0 b2 ðxÞ f2 ðxÞ The integral sliding surface is defined by Zt s ¼ Cðx  x0 Þ  CðA  BKÞxðsÞds ð3:25Þ 0 In (3.3 Integral Sliding Mode Control In 1996. The model to can be rewritten by x_ ¼ Ax þ Bu þ F ð3:24Þ In (3.16). The characteristic of the sliding surface of an ISMC system is that there is an integral term in the surface.24).25). Concerning other SMC methods. the system motion under sliding mode has a dimension short of that of the state space. Under integral sliding mode. the system motion has a dimension equal to that of the state space. i..3. u is the control input. the robustness of the ISMC system against matched uncertainties is promised in the whole state space. where C is selected to guarantee that CB is nonsingular and K is designed via pole assignment such that the eigenvalues of matrix A − BK are less than 0. Accordingly. recall the dynamic model of a nominal crane system (2.3 Integral Sliding Mode Control 79 3. C and K are 1 × 4 vectors. Utkin and Shi [6] proposed an improved sliding control method named integral sliding mode control (ISMC).1 Control Design For the purpose of illustrating the ISMC design. the matrixes A. 3.

29).2 Stability Analysis On the Lyapunov’s stability scheme.29) with respect to time t and substituting the crane model (3. CBueq þ CF þ CBKx ¼ 0.30) can be written by V_ ¼ sðCBusw Þ ð3:31Þ _ VðtÞ  0 in (3. the switching control law is still kept unknown. (3. To obtain the expression of ueq. s_ ¼ 0 exists and the sole equivalent control law ueq is applied to the crane control system. Then. From (3. a Lyapunov candidate function of the ISMC-based crane control system has a form of 1 VðtÞ ¼ s2 ð3:29Þ 2 Differentiating V in (3.31) should exist to make the crane control system asymptotically stable.27).25).26).80 3 Overhead Crane Control by Sliding Mode Methods Adopt the equivalent-control-based SMC design [6]. 3. the integral sliding surface (3. differentiating s in (3. and the equivalent-control-based ISMC law (3.28). VðtÞ\0 in (3.26) into the derivative of V yield V_ ¼ s_s ¼ sðCBu þ CF þ CBKxÞ ð3:30Þ ¼ sðCBueq þ CBusw þ CF þ CBKxÞ From (3.3. In (3. Define the ISMC law as u ¼ ueq þ usw ð3:26Þ where ueq is the equivalent control law and usw is the switching control law.24). As a result.24) into the derivative of s yield s_ ¼ C_x  CðA  BKÞx ¼ CBu þ CF þ CBKx ð3:27Þ When the crane states slide along the integral sliding surface (3.25) with respect to time t and substituting the crane model (3.25). the switching control law will be deduced from the stability analysis of the crane ISMC system. In the sense of Lyapunov. these linguistic descriptions are formulated by ueq ¼ Kx  ðCBÞ1 CF ð3:28Þ The equivalent control law has been obtained. define . To obtain the whole ISMC law.

indicating that the stability of the ISMC system is kept unchanged when the control law (3.28).33) is carried out on crane systems with matched uncertainties (2. uÞT Differentiating V in (3.24). To test the effects of unmatched uncertainties on the system stability.20).3 Integral Sliding Mode Control 81 usw ¼ ðCBÞ1 ½js þ gsgnðsÞ ð3:32Þ where both κ and η are the positive constants.19). Equation (3. and d(t) is determined by dðx.28) into the derivative of V yield V_ ¼ s_s ¼ sðCBu þ CF þ Cd þ CBKxÞ ð3:35Þ ¼ sðCBueq þ CBusw þ CF þ Cd þ CBKxÞ From (3.3. As mentioned. However.28) and (3. the integral sliding surface (3. uÞ ð3:34Þ where A.35) can be simplified by V_ ¼ sðCBusw þ CdÞ ð3:36Þ Define the switching control law as usw ¼ ðCBÞ1 ½k  s þ g sgnðsÞ ð3:37Þ where κ* and η* are the positive constants and η* should be g [ jjCjj1 jjdjj1 ð3:38Þ .32).33).34) can be drawn. Finally. x_ ¼ Ax þ Bu þ F þ dðx. Equation(3. the ISMC law of the crane system can be deduced from (3. the crane model with unmatched uncertainties (2. the invariance property of SMC can resist matched uncertainties.25).20) is taken into considerations. uncertainties exist everywhere.26) and the equivalent control law (3. CBueq þ CF þ CBKx ¼ 0. uÞ 0 n2 ðx. uÞ ¼ ½0 n1 ðx. determined by u ¼ Kx  ðCBÞ1 CF  ðCBÞ1 ½js þ g sgnðsÞ ð3:33Þ The ISMC law for nominal single-pendulum-type overhead cranes has been drawn in (3. B.29) with respect to time t and substituting the crane model with unmatched uncertainties (3. and the equivalent-control-based ISMC law (3. and F are defined in (3. Rearrange the crane model (2.34).

7c.29).7.38). d has no residual oscillations by the integral sliding mode controller because there are no uncertainties in the nominal crane system.1. η* in (3. VðtÞ\0 exists as well. the ISMC laws for nominal and uncertain cranes have been completed.5 1.7a–d. respectively. The sole difference is the amplitude of the signum function.33) and (3.38) into (3.36) yields V_ ¼ s½k  s  g sgnðsÞ þ Cd   k s2  g jsj þ jjCdjj1 s \  k s2  g jsj þ jjCjj1  jjdjj1 s\0 _ VðtÞ  0 is defined in (3.39) not only has the ability to generate two kinds of dynamic equation.28) and (3.7 −6 −2]T. C = [0. the ISMC law can be deduced from (3.39). the trolley velocity. 3. Otherwise.3 Simulations of Single-Pendulum-Type Cranes by ISMC To demonstrate the performance of the ISMC law for nominal single-pendulum-type overhead crane systems. However. Both the expressions almost look the same. η = 0.7 −3 −1]T and K = [1.2 1. 3.8. Other controller parameters in (3. Compare the two control laws (3. The system performance of the single-pendulum-type nominal overhead crane by the ISMC law is shown in Fig.37).7a.82 3 Overhead Crane Control by Sliding Mode Methods Substituting (3. but also helps the control system resist unmatched certainties. Some numerical simu- lation results are demonstrated in Figs. some simulation results are dis- played and some analyses are presented as well. b. obtained by a Simulink model of MATLAB in Appendix C. 3. From Fig. the ISMC system for single-pendulum-type overhead cranes with unmatched uncertainties is of asymptotic stability.3.37) and (3. The payload in Fig. η is employed to generate two kinds of dynamic equations so that it can be arbitrarily small. the payload angle. 3.33) are determined by κ = 10.33). where the trolley position. η* in (3. the trolley arrives at the desired position at about 4 s. the ISMC system for overhead cranes has no guaranteed stability against unmatched certainties. 3. written by u ¼ Kx  ðCBÞ1 CF  ðCBÞ1 ½j s þ g sgnðsÞ ð3:39Þ So far. These controller parameters are tuned after trial and error.01. The system performance indicates that the ISMC law can resist the single-pendulum . and the payload angular velocity are shown in Fig. At last.7 and 3. In (3. For the purpose of comparison. physical parameters of the overhead crane system and the initial and desired states are kept unchanged from Table 3. 3. In the sense of Lyapunov.39) should coincide with the inequality (3.

3. However.8. where the control input and the integral sliding surface are located in Fig. The trolley moves toward the desired position but the payload has an angular deviation because Fig. b Trolley velocity. the driven force f in Fig. respectively.8a.8a is applied to the trolley as largely as possible at the outset. c Payload angle.8 Control performance by the ISMC method for nominal single-pendulum-type overhead crane systems. Many uncertainties exist in a real crane system.7 System performance by the ISMC method for nominal single-pendulum-type overhead crane systems. an accurate dynamic model is hard to obtain. d Payload angular velocity motions very well if the crane model is accurate enough. 3. 3. b Integral sliding surface . To accelerate the trolley as fast as possible. 3. it is meaningful to test the system per- formance in the presence of uncertainties. To make the designed ISMC law practical.3. a Control input. 3. b.3 Integral Sliding Mode Control 83 Fig. The control performance of the closed-loop system is demonstrated in Fig. a Trolley position.

the integral sliding mode in Fig.10a. η* = 0. some slight oscillations in Fig. 3. κ* = κ = 10. Simulation results are shown in Figs.7.10 illustrates the control performance by the ISMC method.38). 3. Since it is hard to accurately model the uncertainties.9 shows the system performance by the ISMC method where the trolley position. η* should coincide with the inequality (3. the parameter vectors C and K of the ISMC sliding surface are kept unchanged from the ISMC system for the nominal overhead crane. Such a deviation is small enough in practice. For the purpose of comparison.1. 3. Compared with the zeroth residual oscillations in Fig. where the control input and the sliding surface are shown in Fig. the payload angle. 3. 3.10 is very similar to that of plotting Figs. respectively. the control input slightly jumps back and forth to resist the adverse effects of unmatched uncertainties. From (3. the ISMC system for the nominal crane enters its integral sliding mode at t = 0 s. although the sliding mode begins at t = 0 s. the state trajectories deviate from the integral sliding surface because of the unmatched uncertainties. Meanwhile. From Fig. 3. u) || = ||ξ2(x.10. From Fig. Figure 3. and the payload angular velocity are shown in Fig.9 indicates that some slight oscillations of the payload are inevitable in a real crane system. Further.8b. 3.6 is set to guarantee the asymptotic stability of the ISMC system for the crane in the presence of uncer- tainties. Once the trajectories are out of the sliding .8b. both the uncertain terms ξ1(x. Figure 3.10a. 3.84 3 Overhead Crane Control by Sliding Mode Methods of its inertia.10b is not ideal because of the existence of uncertainties. The uncertainties always tend to make the state trajec- tories deviate from the sliding surface.9a–d. the payload has some residual oscillations because of the existence of unmatched uncertainties although the crane can arrive at the desired position at about 4 s. u) are determined by 0. The Simulink model of MATLAB to plot Figs. b. ||ξ1(x. The maximum angular deviation is about 0. ||C|| = 3.4 Simulations of Uncertain Single-Pendulum-Type Cranes by ISMC Recall the crane model with unmatched uncertainties (2. 3. Consequently. 3. From Fig. 3. the trolley velocity.1 × [1–2 × rand()]. Thereafter. the trolley decreases its velocity till the control task is achieved.3.9 and 3. u) || = 0.19). According to the parameter tuning. where rand() is a MATLAB command to generate a uniformly distributed pseudorandom number on the open interval (0.7 and 3.9 and 3. 3. In the model.8.12 rad.9. 3. Then. respectively. the driven force f decreases and reverses its direction against the angular deviation. the state trajectories keep sliding on the integral sliding surface and converge to the origin.10b. 1). In Fig. Fig.9 exist in the crane system because of the effects of unmatched uncertainties. u) and ξ2(x. Compared with the ideal integral sliding mode in Fig. 3.38). 3.

c Payload angle.3. b Trolley velocity. . The process con- tinues till the end of the simulation such that the curve of the control input in Fig. 3. the ISMC law draws the trajectories back to the surface.9 System performance by the ISMC method for the single-pendulum-type overhead crane system with unmatched uncertainties.10a does not smooth. d Payload angular velocity Fig. 3.10 Control performance by the ISMC method for the single-pendulum-type overhead crane system with unmatched uncertainties. a Control input. a Trolley position. 3. b Sliding surface mode.3 Integral Sliding Mode Control 85 Fig.

16). the asymptotic stability of the first-order SMC is promised which leads to the convergence of the states to the origin. The model can be rearranged by € ¼ Bu þ F X ð3:40Þ In (3. recall the dynamics of single-pendulum-type over- head crane systems (2.4. the TSM control method will be employed to deal with the crane control problem. continuous stirred tank reactor [10]. . u is the control input. For a positive constant T > 0. 2. b1(x). the vector P(t) is defined by P(t) = [p1 p2]T where p1 and p2 are assumed to satisfy p1 and p2: R+→R. p_ i ð0Þ ¼ e_ i ð0Þ and €pi ð0Þ ¼ €ei ð0Þ here i = 1. But this convergence can just be guaranteed within infinite time. driven force applied to the trolley. b2(x)]T and F = [f1(x). To design the TSM control law for the type of cranes. p1 and p2 2 C2[0. In a TSM control system. for example. 3. In Sect. f1(x). a new type of SMC. The surface s has the form of   s ¼ cT1 ðXe  PðtÞÞ þ cT2 X_ e  PðtÞ _ ð3:41Þ In (3. p2 and p_ 2 2 L∞. and the error vector can be written by Xe = [e1 e2]T.. was invented by Venkataraman and Gulati [7]. Several open questions still remain on the mathematical treatment of the system’s behavior at the origin since it is non-Lipschitz. Evolved out of seminal work on terminal attractors.∞). the main idea of TSM control is evoked by the concept of terminal attractors which guarantees finite-time convergence of the states. f2(x)]T. X = [x θ]T. p1. nonlinear spacecraft attitude [9]. pi is bounded as 0 ≤ t ≤ T and pi ð0Þ ¼ ei ð0Þ. 3. etc. p_ 1 . i. where x. the vectors c1 and c2 2 <21 are constant. and f2(x) are defined in (2.4 Terminal Sliding Mode Control In 1993.40).1 Control Design For the purpose of illustration. B = [b1(x). The TSM control has been widely applied to non- linear process control. a terminal sliding surface should be defined at first. named terminal sliding mode (TSM).86 3 Overhead Crane Control by Sliding Mode Methods 3. After the sliding surface is intercepted.41).e. the error vector Xe is defined by X − Xd where X is the state vector and Xd is the desired state vector. rigid robot control [8].4. b2(x). In contrast. a nonlinear term is introduced in the sliding surface design so that the manifold is formulated as an attractor.16). the trajectory is attracted within the manifold and converges to the origin following a power rule.

ð3:42Þ where the parameters cil (i = 0. 1. l = 0.43). 5 > > T i T > > i : if t[T 0.41) with respect to time t and substituting the crane model (3.45) by the Gaussian elimination algorithm. differentiating s in (3. and (3. Then. 1. p_ i ðtÞ and €pi ðtÞ are zeroth at t = T such that 8 < c00 þ c10 þ c20 ¼ 1 3c þ 4c10 þ 5c20 ¼ 0 ð3:43Þ : 00 6c00 þ 12c10 þ 20c20 ¼ 0 8 < c01 þ c11 þ c21 ¼ 1 3c þ 4c11 þ 5c21 ¼ 1 ð3:44Þ : 01 6c01 þ 12c11 þ 20c21 ¼ 0 8 < c02 þ c12 þ c22 ¼ 0:5 3c þ 4c12 þ 5c22 ¼ 1 ð3:45Þ : 02 6c00 þ 12c12 þ 20c22 ¼ 1 Solve (3.4 Terminal Sliding Mode Control 87 Select pi as 8   > > 1 :: c02 :: > > e ð0Þ þ e_ ð0Þt þ e ð0Þt 2 þ c00 e ð0Þ þ c01 _ e ð0Þ þ e ð0Þ t3 > > i  i 2 i T 3  i T 2 i T i > > > > :: < þ cT104 ei ð0Þ þ cT113 e_ i ð0Þ þ cT122 eð0Þ t4 pi ðtÞ ¼ i > >   if 0tT > > c22 :: > > þ c20 c21 _ 5 ei ð0Þ þ T 4 e ð0Þ þ 3 eð0Þ t . These parameters are obtained and they are listed by 8 8 8 < c00 ¼ 10 < c01 ¼ 6 < c00 ¼ 1:5 c ¼ 15 c ¼8 c ¼ 1:5 ð3:46Þ : 10 : 11 : 10 c20 ¼ 6 c21 ¼ 3 c20 ¼ 0:5 Substitute (3.40) into the derivative of s yield     s_ ðtÞ ¼ c1 X_ e  PðtÞ _ þ c2 X€ e  PðtÞ v     ¼ c1 X_ e  PðtÞ _ þ c2 X€ X € d  PðtÞ v ð3:48Þ     ¼ c1 X_ e  PðtÞ _ þ c2 Bu þ F  X € d  PðtÞ v . From the assumption.46) into (3. the terminal sliding surface can be drawn. (3. To obtain the expression of ueq. Adopt the equivalent-control-based SMC method. 2. the values of pi(t).44).41). 2) can be derived from the twice differentiable assumption. Define the TSM control law as u ¼ ueq þ usw ð3:47Þ where ueq is the equivalent control law and usw is the switching control law.3.

the TSM control law can be deduced from (3.51) yields V_ ¼ sðc2 Busw Þ ð3:52Þ Define the switching control law as usw ¼ ðc2 BÞ1 ½js þ gsgnðsÞ ð3:53Þ where both κ and η are the positive constants. the terminal sliding surface (3.52) can be written as V_ ¼ js  gjsj\0 Since VðtÞ  0 in (3.53).48). the switching control law will be deduced from the system stability analysis. determined by . To obtain the whole TSM control law.2 Stability Analysis To verify the system stability on the Lyapunov’s stability scheme.4. 3.49). the derivative of V in (3. Substituting it into (3. a Lyapunov function is defined by 1 VðtÞ ¼ s2 ð3:50Þ 2 Differentiating V in (3.40). and the equivalent-control-based TSM control law (3.50).47).49) into the derivative of V yield V_ ¼ sðtÞ_sðtÞ      ð3:51Þ ¼ sðtÞ c1 X_ e  PðtÞ _ € d  PðtÞ þ c2 Bðueq þ usw Þ þ F  X € In (3. Then. €  c2 X c2 Bueq þ c2 F  c2 PðtÞ € d þ c1 ðX_ e  PðtÞÞ _ ¼0 exists. the crane control system is asymptotically stable in the sense of Lyapunov. these linguistic descriptions are formulated by    €  c2 X ueq ¼ ðc2 BÞ1 c2 F  c2 PðtÞ € d þ c1 X_ e  PðtÞ _ ð3:49Þ From the TSM control law (3.41). the switching control law is still kept unknown.50) with respect to time t and substituting the crane model (3.88 3 Overhead Crane Control by Sliding Mode Methods When the crane states slide along the terminal sliding surface (3.41).49) and (3. s_ ¼ 0 exists and the sole equivalent control law ueq is applied to the crane control system. From (3. Finally.

19). indicating that the stability of the TSM control system is kept unchanged if the control law (3. the invariance property of SMC can resist matched uncertainties. Equation (3.55) can be drawn. the crane control system on basis of the TSM control law is asymptotically stable in the sense of Lyapunov.57) and (3. Rearrange the uncertain crane model (2. Then.56).3. To test the effects of unmatched uncertainties on the system stability.49) and (3. B.54) is carried out by the crane system with matched uncertainties.47) and the equivalent control law (3. Finally.56) becomes _ VðtÞ ¼ js2  gjsj þ sc2 K\  js2  gjsj þ jjc2 Kjj1 s \  js2  gjsj þ jjc2 jj1 jjKjj1 s\0 Since VðtÞ  0 in (3. and K is determined by K ¼ ½n1 ðx.41).40). the TSM control law for the crane system with unmatched certainties can be deduced from (3. the crane model with unmatched uncertainties (2.4 Terminal Sliding Mode Control 89     €  c2 X u ¼ ðc2 BÞ1 c2 F  c2 PðtÞ € d þ c1 X_ e  PðtÞ _ þ js þ g sgnðsÞ ð3:54Þ As mentioned. and F are defined in (3. uÞT Differentiating V in (3.55).20) is taken into considerations. written by €  c2 X u ¼ ðc2 BÞ1 ½c2 F  c2 PðtÞ € d þ c1 ðX_ e  PðtÞÞ _ þ j s þ g sgnðsÞ ð3:59Þ .49) into the derivative of V yield V_ ¼ sðc2 Busw þ c2 KÞ ð3:56Þ Define the switching control law as usw ¼ ðc2 BÞ1 ½j s þ g sgnðsÞ ð3:57Þ where both κ* and η* are the positive constants and η* should be g [ jjc2 jj1 jjKjj1 ð3:58Þ Substitute (3.57). the terminal sliding surface (3. and the equivalent-control-based TSM control law (3.50) with respect to time t and substituting the crane model with unmatched uncertainties (3. € ¼ Bu þ F þ F X ð3:55Þ where X.58) into (3. uÞ n2 ðx.50). the derivative of V in (3.

11 System performance by the TSM control method for nominal single-pendulum-type cranes. the payload angle. the state trajectories switch Fig. 3. From Fig. respectively. Figure 3. From Fig.90 3 Overhead Crane Control by Sliding Mode Methods 3.1.7c.12a.45]T. T = 1. and η = 0. 3. 3.04. c Payload angle. 3. d has some residual oscillations by the TSM controller.11a. d Payload angular velocity . κ = 1.12. Some numerical simulation results are demonstrated in Figs. and the payload angular velocity are shown in Fig. the curve of the driven force f applied to the trolley has the chattering phenomenon. 3. From Fig. The system performance of the overhead crane is displayed in Fig. physical parameters of the overhead crane system and the initial and desired states are kept unchanged from Table 3. 3.12 shows the control performance of the closed-loop crane control system.5. a Trolley position.54) are determined by c1 = [–1 20]T.4. 3.11a–d. b. 3. b. the trolley arrives at the desired position at about 2 s. c2 = [–15 –3. the trolley velocity. For the purpose of comparison.12a.12b. The payload in Fig. obtained by a Simulink model of MATLAB in Appendix D.11. indicating that the payload slightly sways back and forth around the downward position. respectively. 3. some numerical simulation results are displayed and some analyses will be presented as well.11 and 3. Other controller parameters in (3. where the control input and the sliding surface are located in Fig. where the trolley position. b Trolley velocity.3 Simulations of Single-Pendulum-Type Cranes by TSM To demonstrate the performance of the TSM control law for nominal single-pendulum-type overhead crane systems.

The maximum angular deviation is about 0.14a. 3.13 and 3. where rand() is a MATLAB command to generate a uniformly distributed pseudorandom number on the open interval (0.58). the sliding surface variable s jumps back and forth in the dynamic process. 3. Simulation results are shown in Figs.13a. In the model. 3.19). a Control input. The trolley position. 3.14 is very similar to that of plotting Figs. The payload has slight residual oscillations because of the existence of unmatched uncertainties.13 and 3.12 Control performance by the TSM control method for nominal single-pendulum-type cranes. b Sliding surface around the terminal sliding surface because the nonlinear term P(t) exists in the surface. b. 3.4 Simulations of Uncertain Single-Pendulum-Type Cranes by TSM Recall the crane model with unmatched uncertainties (2. both the uncertain terms ξ1(x. indicating that the chattering phenomenon indeed exists in the closed-loop control system. η* = 2 to guarantee the asymptotic stability of the control system for the uncertain crane because jjc2 jj1 jjKjj1 ¼ 15  0:1\ 2 in (3. However. the trolley velocity. The Simulink model of MATLAB to plot Figs. u) and ξ2(x. 1).4 Terminal Sliding Mode Control 91 Fig. 3. 3.12.38). the payload angle. 3.11 and 3.25 rad (15°).3. For the purpose of comparison. Such a deviation is small enough in practice. respectively.1 × [1–2 × rand()].14 are kept unchanged from the TSM control system for the nominal overhead crane. 3. From (3. Concerning the crane control based on TSM.13 and 3. 3. the curves in Fig. respectively. From Fig.4. The curves in Fig. and the payload angular velocity are located in Fig. c seem smooth. Moreover.13b. The control input and the sliding surface are shown in Fig.14. u) are determined by 0. the control input severely jumps back and forth to resist the adverse effects of unmatched uncertainties. the parameter vectors c1 and c2 of the terminal sliding surface in Figs.13a–d. To overcome the chattering of the . d slightly jump back and forth. κ* = κ = 1. the terminal attractor in the sliding surface may deteriorate the chattering of the control input. 3.14a.

12b.14b. a Control input. The sliding surface curve is neither smooth nor gradual in Fig. a Trolley position.92 3 Overhead Crane Control by Sliding Mode Methods Fig. one possible method is to employ the saturation function instead of the signum function in (3.59). b Trolley velocity. It looks random but bounded because of both the uncertain terms.13 System performance by the TSM control method for uncertain single-pendulum-type cranes. 3. 3. . Compared with Fig. c Payload angle. d Payload angular velocity Fig. 3. the system trajectories severely deviate from the sliding mode in Fig.14b.14 Control performance by the TSM control method for uncertain single-pendulum-type cranes. 3. 3. b Sliding surface control input.

integral. Theoretically. a second-order sliding surface for the control design [11] should be defined by Zt s ¼ a X_ e þ bT Xe þ cT T Xe ðsÞds ð3:60Þ 0 In (3. To develop the second-order SMC law for such a single-pendulum-type overhead crane system.5. the first-order SMC is usually called the SMC. Preserving the main advantages of the first-order SMC.. Since the sliding surface variable s covers the proportional. it can be tread as a PID-type sliding surface.61). 3. The drawback indicates high-frequency vibrations of the controlled system.5 Second-Order Sliding Mode Control 93 3.5. Then. 3. an approach. In Sect.5 Second-Order Sliding Mode Control Mentioned in Sect. has been proposed in order to reduce the chattering phenome- non.3. (3.61) becomes .60) with respect to time t yields   € e þ bT X_ e þ cT Xe ¼ aT X s_ ¼ aT X € X € d þ bT X_ e þ cT Xe ð3:61Þ Substitute (3. the signum function acts on its higher order time derivative. which degrades the system performance and may lead to inherent instability. i. Differentiating s in (3. called higher order SMC. recall the crane mode (3. The problem of higher order SMC is formulated through the differentiation of the sliding variable and is equivalent to the finite-time stabilization of higher order integrator chains. Instead of influencing the first sliding variable time derivative. chattering phenomenon.e. and derivative of the state vector X. the second-order SMC method is taken into consideration because of a trade-off between control accuracy and design complexity. where X = [x θ]T is the state vector and Xd = [xd θd]T is the desired vector composed of the desired trolley position and the desired payload angle.1. and γ 2 <21 are constant the vectors and Xe is defined by X − Xd.1 Control Design For the purpose of illustration. arbitrary order sliding mode controllers can be developed. Compared to the first-order SMC.40) into (3. 3. In spite of its robustness properties. such a resulting controller has a specific disadvantage.60) α. β.40). the chattering effect is removed to higher derivatives so that the higher order SMC method can also achieve a better accuracy with respect to discrete sampling time.

66) can be drawn. The final control law can be obtained by solving the equation. The second-order SMC law has a form of  1 h T : : u_ ¼ aT B  a B þbT B u  bT F  aT F  ð3:67Þ cT X_ e  ks  j_s  gsgnðsÞ þ bT X € d þ aT X €d The second-order SMC law deduced from (3. its Lyapunov candidate cannot only have the sole sliding surface variable.63). and η are the constants and they are positive. Define a novel Lyapunov candidate function as 1 1 V ¼ ks2 þ s_ 2 þ gjsj ð3:68Þ 2 2 .67). Then.63) becomes   _ þ aT Bu_ þ aT F_  aT X €s ¼ aT Bu € d þ bT Bu þ F  X € d þ cT X_ e ð3:64Þ Let €s ¼ ks  j_s  gsgnðsÞ ð3:65Þ where the scalar parameters λ.62).64). differentiating the derivative of s with respect to time t yields   _ þ aT Bu_ þ aT F_  aT X €s ¼ aT Bu € d þ bT X€ X € d þ cT X_ e ð3:63Þ Substitute (3. Then. (3.40) into (3. Substitute (3.2 Stability Analysis To verify the stability of the crane control system under the second-order SMC law (3. 3. some theoretical analyses are presented in the sense of Lyapunov.66). κ.67) is u(0) = 0 because there is no control input at t = 0. Since the second derivative of the sliding surface variable exists in the second-order SMC method.94 3 Overhead Crane Control by Sliding Mode Methods   € e þ bT X_ e þ cT Xe ¼ aT Bu þ F  X s_ ¼ aT X € d þ bT X_ e þ cT Xe ð3:62Þ In (3. (3.5. _ þ aT Bu_ þ aT F_  ks  j_s  gsgnðsÞ ¼ aT Bu   ð3:66Þ € d þ bT Bu þ F  X  aT X € d þ cT X_ e Rearrange (3.65) into (3.67) is a first-order differential equation. The initial condition in (3.

3.5 Second-Order Sliding Mode Control 95

Differentiating V in (3.68) with respect to time t yields

V_ ¼ ks_s þ s_€s þ g_s sgnðsÞ
ð3:69Þ
¼ s_ ½ks þ €s þ g sgnðsÞ

Substituting (3.64) and (3.67) into (3.69) yields

V_ ¼ j_s2 ð3:70Þ

From (3.68), V > 0 exists. Since the derivative of V is negative in (3.70), the
crane system (3.40) by the second-order SMC law (3.67) is of asymptotic stability.
The control law (3.67) can asymptotically stabilize nominal crane systems
(3.40). To test the effects of unmatched uncertainties on the system stability, the
crane model with unmatched uncertainties (3.55) is taken into considerations.
Concerning the second-order SMC law for crane systems in the presence of
unmatched uncertainties, define the sliding surface as (3.60) and differentiating s in
(3.60) with respect to time t. Then, (3.61) can be drawn. Substituting (3.55) into
(3.61) yields
 
€ e þ bT X_ e þ cT Xe ¼ aT Bu þ F þ K  X
s_ ¼ aT X € d þ bT X_ e þ cT Xe ð3:71Þ

In (3.71), differentiating the derivative of s with respect to time t yields
 
_ þ aT Bu_ þ aT F_ þ aT K
€s ¼ aT Bu _  aT X
€ d þ bT X€ X
€ d þ cT X_ e ð3:72Þ

Substitute (3.55) into (3.72). Then, (3.72) becomes
 
_ þ aT Bu_ þ aT F_ þ aT K
€s ¼ aT Bu _  aT X
€ d þ bT Bu þ F þ K  X
€ d þ cT X_ e
ð3:73Þ

Let

€s ¼ ks  j_s  gsgnðsÞ  qsgnð_sÞ ð3:74Þ

where the scalar parameters λ, κ, η, and ρ are the constants and they are positive.
Rearrange (3.73) and (3.74). The second-order SMC law for
single-pendulum-type overhead cranes in the presence of unmatched uncertainties
has a form of
 1 h
T : :
u_ ¼ aT B  a B þbT B u  bT F  aT F
 ð3:75Þ
cT X_ e  ks  j_s  gsgnðsÞ  qsgnð_sÞ þ bT X
€ d þ aT X
€d

96 3 Overhead Crane Control by Sliding Mode Methods

where ρ is subject to

_ þ bT Kjj
q [ jjaT K ð3:76Þ
1

To verify the stability of the crane control system by the second-order SMC law
(3.67), define a Lyapunov candidate function as (3.68), differentiate V with respect
to time t, and substitute (3.73) and (3.75) into the derivative of V. Then, (3.77) can
be drawn.
 
_ þ bT K
V_ ¼ s_ ½ks þ €s þ gsgnðsÞ ¼ s_ j_s  qsgnð_sÞ þ aT K ð3:77Þ

Consider the inequality (3.76). Equation (3.77) can be written by
 
V_ ¼ j_s2  qj_sj þ aT K _ þ bT K s_
ð3:78Þ
_ þ bT Kjj j_sj\0
  j_s2  qj_sj þ jjaT K 1

_
Since VðtÞ  0 in (3.68) and V\0 in (3.78), the second-order SMC law (3.75)
can guarantee that the crane system subject to unmatched uncertainties is asymp-
totically stable in the sense of Lyapunov.

3.5.3 Simulations of Single-Pendulum-Type Cranes
by Second-Order SMC

To demonstrate the performance of the second-order SMC law for nominal
single-pendulum-type overhead crane systems, some numerical simulation results
are displayed and some analyses will be presented as well.
For the purpose of comparison, physical parameters of the overhead crane
system and the initial and desired states are kept unchanged from Table 3.1.
Controller parameters of the second-order SMC law in (3.67) are selected after trial
and error and they are determined by α = [–10 0.5]T, β = [–10 35]T, γ = [0.01 0.01]T,
λ = 1, κ = 1.9, and η = 0.1, respectively. Some numerical simulation results are
demonstrated in Figs. 3.15 and 3.16, obtained by a Simulink model of MATLAB in
Appendix E.
The system performance of the overhead crane is displayed in Fig. 3.15, where
the trolley position, the trolley velocity, the payload angle, and the payload angular
velocity are shown in Fig. 3.15a–d, respectively. From Fig. 3.15a, b, the trolley
arrives at the desired position at about 7 s. Compared with the first-order SMC law,
the integral SMC law, and the terminal SMC law, the second-order SMC law
cannot make the crane system effective or productive because the trolley moves
slowly. However, the maximum angular deviation of the payload is just 0.02 rad
(about 1.15°). Such an angular derivation is the smallest one among the four SMC
methods.

3.5 Second-Order Sliding Mode Control 97

Fig. 3.15 System performance by the second-order SMC method for nominal
single-pendulum-type cranes. a Trolley position; b Trolley velocity; c Payload angle; d Payload
angular velocity

Figure 3.16 shows the control performance of the closed-loop crane control
system, where the control input and the sliding surface are located in Fig. 3.16a, b,
respectively.
As mentioned, the purpose of the second-order SMC law is to reduce the
chattering phenomenon by covering the signum function in the twice derivative of
the sliding surface s. From Fig. 3.16a, the curve of the driven force f applied to the

Fig. 3.16 Control performance by the second-order SMC method for nominal single-
pendulum-typecranes. a Control input; b Sliding surface

98 3 Overhead Crane Control by Sliding Mode Methods

trolley almost has no chattering phenomenon and the design purpose has been
achieved. On the other hand, the maximum control input is just 0.6 N, which is also
the smallest one among the four SMC methods. The fact means that only a small
rated power driver can be employed to achieve the control task in practice. From
Fig. 3.12b, the state trajectories arrive at the second-order sliding surface at 7 s, and
then the system trajectories keep sliding on the surface till the end of the dynamic
process.

3.5.4 Simulations of Uncertain Cranes by Second-Order
SMC

Recall the crane model with unmatched uncertainties (2.19). In the model, both the
uncertain terms ξ1(x, u) and ξ2(x, u) are determined by 0.1 × [1–2 × rand()], where
rand() is a MATLAB command to generate a uniformly distributed pseudorandom
number on the open interval (0, 1). Due to the existence of uncertainties, the control
law again unmatched uncertainties (3.75) is adopted
Compared with (3.67), one more term exists in (3.75) and the term is very
important because it is related to the system stability. For the purpose of compar-
ison, the controller parameters α, β, γ, λ, κ, and η are kept unchanged from the
second-order SMC law (3.67) for the nominal overhead crane. According to (3.76),
ρ = 2.5 are selected to guarantee the asymptotic stability of the control system for
the uncertain crane because of jjaT K _ þ bT Kjj \2:5. Some numerical simulation
1
results are shown in Figs. 3.17 and 3.18. The Simulink model of MATLAB to plot
Figs. 3.17 and 3.18 is very similar to that of plotting Figs. 3.15 and 3.16.
The trolley position, the trolley velocity, the payload angle, and the payload
angular velocity are located in Fig. 3.17a–d, respectively. The curve in Fig. 3.17a
seems smooth. However, the curves in Fig. 3.13b–d slightly jump back and forth,
indicating that the chattering phenomenon indeed exists in the closed-loop control
system. The payload has slight residual oscillations because of the existence of
unmatched uncertainties. The maximum angular deviation is about 0.02 rad (1.15°).
Such a deviation is small enough in practice.
From Fig. 3.17a, the trolley arrives at the desired position at about 20 s. Such a
slow transport definitely reduces the throughput of the crane system in reality.
However, the payload endures the minimum angular deviation. The contradiction
between the trolley positioning and the payload swing is completely illustrated by
the second-order SMC method. Such a method is a good technique to resist the
payload oscillations, but it fails to a highly effective transport.
The control input and the sliding surface are displayed in Fig. 3.18a, b,
respectively. From Fig. 3.18a, the control input jumps back and forth to resist the
adverse effects of unmatched uncertainties. But the chattering is not very severe
because only a small control input is needed to make the trolley slowly move.

17 System performance by the second-order SMC method for uncertain single-pendulum-type cranes. b Trolley velocity. 3.18 Control performance by the second-order SMC method for uncertain single-pendulum-type cranes. d Payload angular velocity Fig.3.5 Second-Order Sliding Mode Control 99 Fig. a Trolley position. a Control input. c Payload angle. b Sliding surface . 3.

str = []. sizes. otherwise error(['Unhandled flag = '.NumInputs = 4.DirFeedthrough = 0. case 0.u).1 and 3.9} sys=[].NumContStates = 4.u) .NumOutputs = 4.u).4.x0.x0.NumSampleTimes = 0.num2str(flag)]). case 1. sizes.ts]=mdlInitializeSizes.x.str. sizes.NumDiscStates = 0.u. 3.flag) switch flag.100 3 Overhead Crane Control by Sliding Mode Methods Appendices A Simulink Model to Plot Figs. [sys. end function [sys. sizes. sizes.ts] = sfuntmpl(t. % at least one sample time is needed sys = simsizes(sizes).x.ts = [].str. m function [sys. function sys=mdlDerivatives(t. sys=mdlDerivatives(t.ts]=mdlInitializeSizes sizes = simsizes. sizes.x. case 3.x. sys=mdlOutputs(t.2 Plant program: SPCrane.str. x0 = [0 0 0 0]. case {2.x0.

case 0.NumInputs = 4. B=(m*l*cos(x(3))). sizes.305. sizes.8.u. sys(3)=x(4).x.num2str(flag)]).DirFeedthrough = 1. Controller program: FirstOrderSMC. sizes.2.u) sys=x.M=1.flag) switch flag.9} sys=[]. % at least one sample time is needed sys = simsizes(sizes).x0. [sys.l=0.Appendices 101 %parameter g=9. sys(1)=x(2).str. case {1.M=1.x0.ts]=mdlInitializeSizes. C=(cos(u(3))).u). A=(M+m). m function [sys.ts = [].x0.x. C=(cos(x(3))). sizes.m=0.NumDiscStates = 0.l=0.4. . otherwise error(['Unhandled flag = '.81.NumSampleTimes = 0.ts]=mdlInitializeSizes sizes = simsizes.x.ts] = sfuntmpl(t. sizes. sys=mdlOutputs(t. function sys=mdlOutputs(t.8.u) % parameter of system g=9. end function [sys.m=0. B=(m*l*cos(u(3))).81. case 3. sizes. x0 = [].NumContStates = 0. sys(4)=((u(1)+m*l*x(4)^2*sin(x(3)))*C+g*A*sin(x(3)))*inv(B*C-A*l). A=(M+m). function sys=mdlOutputs(t.str. sys(2)=((u(1)+m*l*x(4)^2*sin(x(3)))*l+g*B*sin(x(3)))*inv(A*l-B*C).str.str = [].305. D=(m*l*sin(u(3))).NumOutputs = 4.x.

3)*u(4)+c(1. xlabel({'Time(s)'.4)*b2).2)*f1+c(1.'(b)'}).2.hold on. subplot(2.4)*f2)/(c(1.-3.3. e3=u(3).4)*b2). sys(2)=s. ylabel('x(m)').simout(:.1)*u(2)+c(1.'(a)'}).%angular velocity plot(tout.3)).4)*e4.2. Plot program: crane_plot.1)*e1+c(1.2. xlabel({'Time(s)'. %parameter in control law ita=0.'(c)'}).4)*b2). sys(4)=-ds/(c(1.1). con_law=-(c(1.4)*f2+ds)/(c(1. sys(1)=con_law.2.2).2)*f1+c(1.2)*b1+c(1. f2=(D*C*(u(4))^2+A*g*sin(u(3)))/(B*C-A*l).2)*b1+c(1. s=c(1. e1=(u(1))-1.simout(:. ylabel('w(rad/s)'). xlabel({'Time(s)'.2)).hold on. k=4.3)*u(4) +c(1.simout(:.1]. c=[-3.1)).'(d)'}). xlabel({'Time(s)'. e2=u(2). f1=(D*(u(4))^2+C*D*g/l)/(A-B*C/l). b1=1/(A-B*C/l).%theta plot(tout.2)*e2+c(1. ylabel('\theta(rad)').1)*u(2)+c(1. . b2=C/(B*C-A*l). %displacement of trolley plot(tout.05.%velocity of trolley plot(tout.3).simout(:. ylabel('v(m/s)'). e4=u(4). m figure(1) subplot(2. sys(3)=-(c(1. subplot(2.4)). subplot(2.hold on.hold on.2)*b1+c(1.4).10.102 3 Overhead Crane Control by Sliding Mode Methods %state error.3)*e3+c(1. ds=ita*sign(s)+k*s.

%theta plot(tout. B Simulink Model to Plot Figs. xlabel({'Time(s)'. subplot(2.'(c)'}). subplot(2.5 and 3. 3. xlabel({'Time(s)'.1)).4)).%angular velocity plot(tout.2.2.simout1(:.simout1(:.'(a)'}). xlabel({'Time(s)'.3)). ylabel('sliding surface s').hold on.2. ylabel('Switching control'). ylabel('f(N)').1) plot(tout.4). xlabel({'Time(s)'.2)).simout1(:.hold on. subplot(2.%velocity of trolley plot(tout.6 .'(b)'}).3).2).2.Appendices 103 figure(2) subplot(2.simout1(:.hold on.hold on.'(d)'}). ylabel('Equivalent control').

l1=3. sizes.b=0.3. case {2. dphi=x(4)..104 3 Overhead Crane Control by Sliding Mode Methods Plant program: DPCrane. sys=mdlDerivatives(t. [sys.mh=10.4...u).. case 1. theta=x(5). function sys=mdlDerivatives(t. 0 0 mc*l1*l2*sin(phi-theta)*dtheta.num2str(flag)])..u) mt=50.phi. case 3.1).ts]=plant(t.0.1).. sys(2)=ddq(1.str.DirFeedthrough=1. dtheta=x(6).9}.str.0.0.mc=2. -mc*l2*cos(theta) mc*l1*l2*cos(phi-theta) mc*l2^2].81.dtheta]. x_=x(1).1).ts=[]. sys(1)=x(2). C=[b (mh+mc)*l1*sin(phi)*dphi mc*l2*sin(theta)*dtheta. G=[0. sys=simsizes(sizes). sizes. sizes.str.flag) switch flag.l2=0.x.0]. x0=[0..dphi. dx=x(2).0.theta]. -(mh+mc)*l1*cos(phi) (mh+mc)*l1^2 mc*l1*l2*cos(phi-theta).x. ddx=ddq(1.ts]=mdlInitializeSizes. case 0. ddphi=ddq(2.NumInputs=1. str=[]. .0].1). ddq=inv(M)*(U-G-C*dq).g=9. m function[sys. sizes.mc*g*l2*sin(theta)].x. U=[tol....NumDiscStates=0.(mh+mc)*g*l1*sin(phi).u). sys=[]. dq=[dx. phi=x(3).NumSampleTimes=0.0. q=[x_.ts]=mdlInitializeSizes sizes=simsizes.NumContStates=6. otherwise error(['Unhandled flag='. 0 -mc*l1*l2*sin(phi-theta)*dphi 0]. end function[sys..x0. sizes.u.x0. ddtheta=ddq(3.x0. M=[mt+mh+mc -(mh+mc)*l1*cos(phi) -mc*l2*cos(theta). tol=u(1).x. sizes.NumOutputs=8.. sys=mdlOutputs(t.

sys(3)=phi.b=0. x_=x(1).mc=2. sys(4)=ddq(2..... sys(7)=dtheta. -(mh+mc)*l1*cos(phi) (mh+mc)*l1^2 mc*l1*l2*cos(phi-theta). sys(2)=dx. ddq=inv(M)*(U-G-C*dq)... dtheta=x(6). dq=[dx.9}.x. ddphi=ddq(2.u)..x0.mc*g*l2*sin(theta)]. tol=u(1).1). q=[x_.str.1). phi=x(3).2. M=[mt+mh+mc -(mh+mc)*l1*cos(phi) -mc*l2*cos(theta).81. U=[tol..dtheta]. case {1.x0. dphi=x(4).dphi. sys(6)=theta.ts]=mdlInitializeSizes.l2=0.(mh+mc)*g*l1*sin(phi)..0].0. 0 -mc*l1*l2*sin(phi-theta)*dphi 0].ts]=mdlInitializeSizes sizes=simsizes. otherwise error(['Unhandled flag='. sys(6)=ddq(3.. 0 0 mc*l1*l2*sin(phi-theta)*dtheta.Appendices 105 sys(3)=x(4).l1=3.1).4. -mc*l2*cos(theta) mc*l1*l2*cos(phi-theta) mc*l2^2].str.phi.str.u) mt=50.x0. ddx=ddq(1. sys(4)=dphi.. theta=x(5).g=9. G=[0. dx=x(2).flag) switch flag.ts]=Controller(t. case 3. sys(5)=ddphi. sys=mdlOutputs(t. function sys=mdlOutputs(t.mh=10..u.1). sys=[]. end function[sys. sys(5)=x(6).num2str(flag)]). ddtheta=ddq(3.x.1). sys(1)=x_. [sys. C=[b (mh+mc)*l1*sin(phi)*dphi mc*l2*sin(theta)*dtheta. m function[sys. case 0. . sys(8)=ddtheta. Controller program: FirstOrderSMC_Reduced.3.x.theta].

4). K=70.1).ts=[]. sizes.1). ylabel('\omega_2 (rad/s)'). sizes. ylabel('\omega_1 (rad/s)').106 3 Overhead Crane Control by Sliding Mode Methods sizes.81. subplot(3.'(d)'}). xlabel({'Time(s)'.DirFeedthrough=1. subplot(3.. ylabel('v (m/s)'). sizes.NumDiscStates=0.hold on.5.theta(:.plot(s(:.mh=10..1). xd=4. xlabel({'Time(s)'.x. function sys=mdlOutputs(t.NumOutputs=2. e=x_-xd. ylabel('\theta_2 (rad)').3).1). sys(1)=tol.x(:. ylabel('u (N)'). -(mt+mh+mc)*(lambda*dx+alpha*dphi+beta*dtheta)-K*sat(s).hold on. subplot(3.lambda=0..2)). theta=u(6).2). plot(dphi(:.'(a)'}).2.NumContStates=0. plot(dx(:.mc=2. Plot program: crane_plot.hold on. (mh+mc)*l1*sin(phi)*(dphi)^2+mc*l2*sin(theta)*(dtheta)^2.2)).1).u) x_=u(1). ylabel('x(m)'). s=de+lambda*e+alpha*phi+beta*theta. figure(2) subplot(1.dphi(:. ddphi=u(5). sys(2)=s.l1=3.2.NumInputs=8.2)).1). m figure(1) subplot(3.'(c)'}).2.'(f)'}).3. subplot(3.6). sys=simsizes(sizes).1).'(b)'}).xd=4.str=[]. phi=u(3).s(:.2)). mt=50. sizes. x0=[].hold on. plot(phi(:.dx(:.hold on.plot(u(:.l2=0. dx=u(2). de=dx.NumSampleTimes=0.g=9. plot(x(:.phi(:. plot(theta(:. xlabel({'Time(s)'..u(:.hold on. .'(a)'}). ddtheta=u(8). xlabel({'Time(s)'.2)).1). subplot(3.dtheta(:.2.2)).2. xlabel({'Time(s)'..2)). xlabel({'Time(s)'.'(e)'}).2)).1).2). dphi=u(4).2. ylabel('\theta_1 (rad)'). xlabel({'Time(s)'.alpha=17.5). tol=-(mh+mc)*l1*cos(phi)*ddphi-mc*l2*cos(theta)*ddtheta+b*dx+. sizes. xlabel({'Time(s)'.b=0. subplot(1. ylabel('s').2.2.'(b)'}).1). plot(dtheta(:.hold on.hold on. dtheta=u(7).beta=-11.

ts = [].ts]=mdlInitializeSizes.x0.u).u. .9} sys=[].ts] = sfuntmpl(t. % at least one sample time is needed sys = simsizes(sizes).str.x0. sys=mdlOutputs(t.NumOutputs = 6.NumSampleTimes = 0. It is similar as one in Appendix A Controller program: ISMC. [sys. sizes. case 3. 3. sizes.x.Appendices 107 C Simulink Model to Plot Figs. case 0.x. end function [sys. otherwise error(['Unhandled flag = '.DirFeedthrough = 1. sizes. x0 = []. m. sizes.flag) switch flag. case {2.str.x0.NumInputs = 8. sizes.num2str(flag)]).7 and 3.NumContStates = 0.4.ts]=mdlInitializeSizes sizes = simsizes. str = [].8 Plant program: SPCrane.str. sizes. m function [sys.NumDiscStates = 0.

sys(4)=int(3). f2=(d*c*(u(4))^2+a*g*sin(u(3)))/(b*c-a*l). e=X-R.1/(a-b*c/l). m=0. xlabel({'Time(s)'.ylabel('x(m)'). ita=0.2. s=C*(e-E). sys(2)=int(1).1)). sys(5)=int(4). d=(m*l*sin(u(3))).ddr. M=1.ddtheta=0.dr=0. a=(M+m).x. l=0.0.ddtheta].u(4)].108 3 Overhead Crane Control by Sliding Mode Methods function sys=mdlOutputs(t.u(2). dR=[dr.simout(:.81.hold on. E=[u(5).dtheta=0. B=[0. f1=(a*(u(4))^2+c*d*g/l)/(a-b*c/l). A=[0 1 0 0. subplot(2. %desired state r=1. us=inv(C*B)*(-ita*sign(s)-k*s-C*B*K*e-C*F).7 -3 -1]. R=[r. %stste error X=[u(1). xlabel({'Time(s)'.0 0 0 0].5 1.2).ddr=0.'(a)'}). b=(m*l*cos(u(3))).u(6). int=(A-B*K)*e.dtheta]. theta=0. sys(1)=us.2 1. %displacement of trolley plot(tout. F=[0.2.305.2)).u(3). C=[0.theta.0 0 0 1.simout(:. sys(6)=s. %parameter of controller K=[1. m figure(1) subplot(2.f2].hold on.7 -6 -2].0 0 0 0.u(8)].1). c=(cos(u(3))).%velocity of trolley plot(tout.8.01.0. sys(3)=int(2).u(7).dr. Plot program: crane_plot.ylabel('v(m/s)').f1.u) % parameter of system g=9.k=10.c/(b*c-a*l)]. .'(b)'}).dtheta.

simout1).%angular velocity plot(tout.12 .11 and 3.'(d)'}).simout2). xlabel({'Time(s)'.ylabel('\omega (rad/s)').2.2).plot(tout. plot(tout.3).Appendices 109 subplot(2. xlabel({'Time(s)'. figure(2) %control force and sliding surface subplot(1.1). subplot(2. D Simulink Model to Plot Figs.ylabel('s').hold on.4)).2. xlabel({'Time(s)'.2.3)).'(c)'}).2.%theta plot(tout. subplot(1.ylabel('u(N)').4).simout(:. xlabel({'Time(s)'.hold on.hold on.'(b)'}).hold on. 3.ylabel('\theta(rad)').'(a)'}).simout(:.

C=(cos(u(3))). m function [sys.ts] = sfuntmpl(t. function sys=mdlOutputs(t. It is similar as one in Appendix A Controller program: TSMC.8. sizes.ts]=mdlInitializeSizes.u). % parameter of controller c1=[-1 20]. T=1. if t==0 . case {1. end function [sys.NumInputs = 8. case 0.2. dde2=u(8).ts = [].NumContStates = 0. de2=u(4). E=(m*l*sin(u(3))).e2]. B=(m*l*cos(u(3))). e1=(u(1))-1.str. sizes.str = [].x0.NumDiscStates = 0. K=0. sizes.NumOutputs = 2.l=0.x.81.x0.x0.str. [sys.m=0.305. m.9} sys=[].x. e=[e1. otherwise error(['Unhandled flag = '. D=l.110 3 Overhead Crane Control by Sliding Mode Methods Plant program: SPCrane.M=1. %e=x-xd. case 3. de1=u(2). sizes.DirFeedthrough = 1.flag) switch flag.ts]=mdlInitializeSizes sizes = simsizes.NumSampleTimes = 0.x.45]. % at least one sample time is needed sys = simsizes(sizes). sys=mdlOutputs(t. dde1=u(6). A=(M+m). x0 = [].4.num2str(flag)]). e2=u(3).u) persistent e10 de10 dde10 e20 de20 dde20 % parameter of system g=9. sizes.str.u.04. sizes.c2=[-15 -3.5.

end if t<=T A0=-10/T^3*e10-6/T^2*de10-1.'(a)'}).ddp2]+c1*[de1-dp1.5/T*dde20.p2=0.simout1(:.5/T*dde10. FX=[(E*(u(3))^2+m*g*sin(u(3))*C)/(A- B*C/D).ylabel('u(N)'). sys(2)=s.e2]-[p1.hold on.5/T^2*dde10.dp2=0. de10=u(2). B2=-6/T^5*e20-3/T^4*de20-0.ylabel('w(rad/s)').'(b)'}).ddp2=0.Appendices 111 e10=(u(1))-1.simout(:.5/T^2*dde20. e20=(u(3)).ylabel('x(m)').simout(:. de20=u(4). subplot(2.2).p2])+c2*([de1. sys(1)=con_law. m figure(1) subplot(2. .2. else p1=0.ddp1=0.'(c)'}).dp2]).%angular velocity plot(t. figure(2) %control force subplot(1. p2=e20+de20*t+1/2*dde20*t^2+B0*t^3+B1*t^4+B2*t^5.ylabel('v(m/s)').3)).1)). xlabel({'Time(s)'.'(d)'}). BX=[1/(A-B*C/D). xlabel({'Time(s)'.%theta plot(t.2. B1=15/T^4*e20+8/T^3*de20+1. xlabel({'Time(s)'.2.2).hold on.hold on.hold on. %displacement of trolley plot(t. subplot(1. subplot(2.2.5/T^3*dde20.simout(:.de2]-[dp1.2)).simout(:. dde20=u(8).4)). A1=15/T^4*e10+8/T^3*de10+1. B0=-10/T^3*e20-6/T^2*de20-1.simout1(:.ylabel('\theta(rad)').hold on.2.plot(t.'(b)'}).'(a)'}).dp1=0. dde10=u(6).4).%velocity of trolley plot(t. end s=c1*([e1.2)). Plot program: crane_plot.2.C/(B*C-A*D)]. ddp1=dde10+A0*t*6+A1*t^2*12+A2*t^3*20.hold on. dp1=de10+dde10*t+A0*t^2*3+A1*t^3*4+A2*t^4*5.de2- dp2]+K*sign(s)). A2=-6/T^5*e10-3/T^4*de10-0.plot(t. xlabel({'Time(s)'.1). xlabel({'Time(s)'.5/T^3*dde10. con_law=-inv(c2*BX)*(c2*FX-c2*[ddp1. ddp2=dde20+B0*t*6+B1*t^2*12+B2*t^3*20. subplot(2. dp2=de20+dde20*t+B0*t^2*3+B1*t^3*4+B2*t^4*5. %sliding surface xlabel({'Time(s)'.ylabel('s'). p1=e10+de10*t+1/2*dde10*t^2+A0*t^3+A1*t^4+A2*t^5.3).(E*(u(3))^2*C+A*g*sin(u(3)))/(B*C-A*D)].1)).1).

x0.NumSampleTimes = 0.15 and 3.NumOutputs = 8. sizes. case 0.DirFeedthrough = 0.ts]=mdlInitializeSizes. [sys.NumContStates = 4.u.u). sizes.ts]=mdlInitializeSizes sizes = simsizes. sys=mdlOutputs(t. otherwise error(['Unhandled flag = '.ts = [].x. % at least one sample time is needed sys = simsizes(sizes). sizes.str.4. sys=mdlDerivatives(t. function [sys.flag) switch flag.112 3 Overhead Crane Control by Sliding Mode Methods E Simulink Model to Plot Figs.9} sys=[]. case {2.num2str(flag)]). m.x0.x0.x. sizes.str = []. sizes. case 1. sizes. x0 = [0 0 0 0].ts] = sfuntmpl(t. end function [sys.16 Plant program: SPCrane. case 3.u).NumInputs = 2. .x. 3.str.NumDiscStates = 0.str.

B=(m*l*cos(x(3))).9} sys=[]. sys(4)=((u(1)+m*l*x(4)^2*sin(x(3)))*C+g*A*sin(x(3)))*inv(B*C-A*l).M=1.str.num2str(flag)]).l=0.DirFeedthrough = 1.flag) switch flag.l=0. sys(2)=((u(1)+m*l*x(4)^2*sin(x(3)))*l+g*B*sin(x(3)))*inv(A*l-B*C).u.x0. sizes. case 1. sys=mdlOutputs(t.NumInputs = 10. Controller program: SecOrderSMC. % at least one sample time is needed sys = simsizes(sizes).u) .NumSampleTimes = 1. [sys.305.ts = [0 0].sys(7)=f1. b2=C/(B*C-A*l). f1=(D*(x(4))^2+C*D*g/l)/(A-B*C/l). function sys=mdlOutputs(t. sizes.81. case 3.ts]=mdlInitializeSizes.x.8.NumContStates = 3.C=(cos(x(3))).sys(8)=f2.4.D=(m*l*sin(x(3))).x0. b1=1/(A-B*C/l). case {1.u).8.m=0.u). end function [sys.x. A=(M+m).sys(5)=b1.ts]=mdlInitializeSizes sizes = simsizes. m function [sys.m=0. function sys=mdlDerivatives(t. f2=(D*C*(x(4))^2+A*g*sin(x(3)))/(B*C-A*l).x0.2. sizes.sys(3)=x(3).C=(cos(x(3))).x.sys(2)=x(2).sys(4)=x(4).NumDiscStates = 0.B=(m*l*cos(x(3))).str.x. sizes. sys(1)=x(1). sys(3)=x(4). sys=mdlDerivatives(t.u) g=9.305.ts] = sfuntmpl(t. otherwise error(['Unhandled flag = '.M=1.u) %parameter g=9.str.NumOutputs = 2.81. sys(1)=x(2).x.x.Appendices 113 function sys=mdlDerivatives(t. A=(M+m).str = []. case 0. sys(6)=b2. sizes.4]. sizes. x0 = [0 0 0.

ylabel('s').xlabel({'Time(s)'.01].2.ylabel('x(m)'). alpha=[-10 0.4)). .2.9.5].hold on.0]+beta*[f1.e4]. figure(2) %control force subplot(1.beta=[-10 35]. sys(1)=con_law.%angular velocity plot(t. %sliding surface plot(t. sys(3)=du.b2])*x(3)+alpha*[u(7). subplot(2. e3=u(3).simout(:.'(b)'}).D=(m*l*sin(u(3))).m=0. xlabel({'Time(s)'.hold on. lamda=1.k=1.ylabel('v(m/s)'). con_law=x(3).2.2).hold on.de=[e2. s=alpha*de+beta*e+gama*[x(1). subplot(1.01 0. Plot program: crane_plot. function sys=mdlOutputs(t.e2=u(2).ylabel('f(N)').e2=u(2).8.C=(cos(u(3))).simout(:.u(6)]… +beta*[b1.simout(:. sys(2)=s.xlabel({'Time(s)'.2)).gama=[0.B=(m*l*cos(u(3))).%velocity of trolley plot(t. sys(2)=e3.ylabel('\theta(rad)').1)).e4=u(4).xlabel({'Time(s)'. subplot(2. A=(M+m).e3].de=[e2. e1=u(1)-1.1). -alpha*[0.e3].01 0.2.1).305.b2])*(lamda*s+(alpha*[u(5).81. du=-inv(alpha*[b1.e4=u(4).f2]-beta*[0.hold on.plot(t.ita=0.simout1(:. g=9.u) %state error. e3=u(3).2.. sys(1)=e1. b2=C/(B*C-A*l).0]+gama*de+k*u(10)+ita*sign(s)).'(a)'}).hold on. b1=1/(A-B*C/l). subplot(2.%theta plot(t.u(8)].2).5].M=1.simout(:.xlabel({'Time(s)'.'(c)'}). e1=u(1)-1. e=[e1.hold on.x(2)].3). alpha=[-10 0.l=0.1)).'(d)'}).2)).e4].2.beta=[-10 35]. m figure(1) subplot(2.gama=[0. s=alpha*de+beta*e+gama*[x(1).ylabel('w(rad/s)').114 3 Overhead Crane Control by Sliding Mode Methods %state error.4).xlabel({'Time(s)'.1. e=[e1. %displacement of trolley plot(t..'(b)'}).'(a)'}). f2=(D*C*(u(4))^2+A*g*sin(u(3)))/(B*C-A*l).x. f1=(D*(u(4))^2+C*D*g/l)/(A-B*C/l).01].3)).x(2)].simout1(:.

References 115 References 1. In: Proceedings of 8th Asian control conference. Fang CC. Masoud ZN (2003) Dynamics and control of cranes: a review. Zhao D. pp 602–607 . In: Proceedings of 35th IEEE conference on decision and control. Venkataraman S. Li HX. Yi J. Wang W. Tuan LA. Taiwan. Song ZK. J Mech Sci Technol 27(6):1863–1873 6. Dubbeldam J (2015) Terminal sliding mode control for continuous stirred tank reactor. Liu D. Kaohsiung. Zhu QM. IEEE Trans Autom Control 39(12):2464–2469 9. Springer. ISA Trans 53(1):117–124 10. Utkin V. Gulati S (1993) Control of nonlinear systems using terminal sliding modes. Berlin 4. Liu D (2004) Design of a stable sliding-mode controller for a class of second-order underactuated systems. Chiang HK. Sun KB (2014) Finite-time control for nonlinear spacecraft attitude based on terminal sliding mode technique. Zhao DY. Kobe. J Dyn Syst Meas Control Trans ASME 115(3):554–560 8. Utkin VI (1992) Sliding modes in control and optimization. Nayfeh AH. Guo W (2013) Nonlinear controller design for the underactuated crane system. J Vib Control 9(7):863–908 2. Wu HR (1994) A robust MIMO terminal sliding mode control scheme for rigid robotic manipulators. IEE Proc Control Theory Appl 151(6):683–690 5. Int J Control Autom 6(6):93–104 3. pp 4591–4596 7. Lee SG (2013) Sliding mode controls of double-pendulum crane systems. Chem Eng Res Des 94:266–274 11. Lin WB. Shi J (1996) Integral sliding mode in systems operating under uncertainty conditions. Abdel-Rahman EM. Man Z. Chen GW (2011) Second-order sliding mode control for a magnetic levitation system. Japan. Paplinski AP.

an overhead crane is adopted as benchmark and each control method is carried out by the transport control problem of the crane. adaptive control based on hierarchical sliding surfaces is taken into considerations.Chapter 4 Overhead Crane Control by Hierarchical Sliding Mode Abstract The methodology of hierarchical sliding mode control (HSMC) covers several design methods. ‘below’. The highest level is the organizer which supervises the performance of the overall system.1007/978-3-662-48417-3_4 . i. This chapter proceeds with source codes of all the simulations. categories. aggregated SMC. Qian and J. Fort purpose of illustration. etc. At first. incremental SMC. © Springer-Verlag Berlin Heidelberg 2016 117 D. Hierarchical Sliding Mode Control for Under-actuated Cranes. is dominated by typical hardware control methods. the HSMC design is also extended for double-pendulum-type overhead cranes. aimed for end-point control tasks. A hierarchical control system is a form of control systems in which a set of devices and governing software are arranged in a hierarchical tree. Yi.) in which the items are represented as being ‘above’. Pointed out by Saridis [1].e. The next level is that of a coordinator which utilizes new computer architectures to effectively control the overall hardware system. Then. DOI 10. and combining SMC. names.. In addition.1 Problem Description A hierarchy is an arrangement of items (objects.   Keywords Overhead crane Hierarchical sliding surfaces Aggregated SMC  Incremental SMC  Combining SMC and adaptive control based hierarchical  sliding surfaces Transport control 4. The lowest level. this chapter introduces these basic HSMC meth- ods. the hierarchical method consists of a three-layer control of “increasing intelligence and decreasing precision.” The lowest level consists of several controllers designed for effective control with existing hardware using an approximation theory of optimal control. values. Both highest levels are computer implemented and the research involved is in developing the appropriate architecture and software to accommodate others. or ‘at the same level as’ one another.

where each sub- system is described by two state variables. incremental hier- archical sliding surfaces...1. respectively. According to the hierarchical theory of control systems [1]. the HSMC method focuses on the hierar- chical structure of the sliding surfaces and designs the control law on basis of the hierarchical sliding surfaces. recall the system state equations (2.118 4 Overhead Crane Control by Hierarchical Sliding Mode What is the relationship between hierarchical sliding mode control (HSMC) and hierarchical control system? Concerning the methodology of HSMC [2.2 Aggregated HSMC Recall the crane structure in Fig.2. and combining hierarchical sliding surfaces. Recall the state space model of single-pendulum-type overhead cranes (2. Two sliding surfaces can be designed for the two subsystems at first. trolley and payload. These design methods treat the four state variables as a whole such that the sliding surfaces have a single-layer structure. 3. There are four state variables in (2. Different sequences of combination can result in different types of hierarchical structures.16). Further. Another way to describe the degree of importance of each state is by structure design of sliding surfaces. which will be introduced in Sects. The basic idea behind the novel structure is as follows [3]. There are four state variables in (2. each state variable plays a unique role in the system performance. aggregated hierarchical sliding surfaces. Some may contribute more to the system performance than others. adaptive control design based on hierarchical sliding surfaces is also investigated in this chapter. These state variables are not equal to each other and they have different degrees of importance.3 and 4. The first-order sliding surface weights each state variable by its coefficient. 2. To develop a diversity of the hierarchical sliding-surface structures for crane control. A hierarchical sliding mode controller is located at the lowest level. the hierarchy means the hierarchical structure of the sliding surfaces instead of the hierarchy of the SMC system. This is our motivation to develop the methodology of HSMC. On the other hand. Briefly. 4. However. So far. the four state variables can arbitrarily combine to generate sliding surfaces. a sliding surface covers several state variables. which does not concern the two higher levels. i. a novel type of sliding surface structure can be constructed according to the mechanical structure of the type of cranes. the HSMC method investigates the sliding mode controller design. a crane dynamic model should be taken into considerations. Then the two surfaces can be . In Chap. Intuitively.16).4. A first-order sliding surface is constructed by linear combination of the state variables. 4. The single-pendulum-type overhead crane is composed of two subsystems.16). 3].e. As a result. 4. the state vector is directly employed for the control designs of several typical sliding mode methods. i.16) and they consist of a state vector. there are three typical types of hierarchical structures of sliding surfaces. Besides the typical hierarchical structures.e.

4. The higher sliding surface is constructed by the two subsystem sliding surfaces. Apparently. Although there are four state variables in the single-layer sliding surface of the traditional first-order SMC. To illustrate and verify the aggregated HSMC for overhead cranes.16). Consequently. i. the novel structure of sliding surfaces is of hierarchy. aggregate the two sliding surfaces and design the second layer sliding surface as S ¼ as1 þ s2 ð4:3Þ . The sliding surface of each subsystem contains two state variables. the aggregated structure is not a simple combination of the four state variables. the control design. Further.1 Control Design Recall the crane model (2. the HSMC design exhibit how the control function affect the motions of the trolley and payload.1) and (4. Consider the traditional first-order SMC. such an aggregated structure indi- cates that there are four state variables in the higher sliding surface. equivalent control law and switching control law. the stability analysis.2) are located at the first layer. both the subsystem sliding surfaces (4. Consider the equivalent-control-based design of the traditional first-order SMC.2. there are two parts in the SMC law. Such design cannot answer the question “How does the control input affect the motions of trolley and payload?” In contrast to the traditional first-order SMC. the main improvement of the aggregated HSMC is its control design. The HSMC design on basis of the aggre- gated sliding-surface structure considers the sliding mode of each subsystem on each subsystem sliding surface and constructs the HSMC law to guarantee the stability of the hierarchical sliding surfaces. consider the four state variables of the crane.e. Compared with the traditional structure. The design is universal and feasible but it fails to exhibit the physical nature of single-pendulum-type overhead cranes. Design the sliding surfaces of the trolley and payload subsystems as s 1 ¼ c1 x1 þ x2 ð4:1Þ s 2 ¼ c2 x3 þ x4 ð4:2Þ Here c1 and c2 are positive constants in order to have the stability of each subsystem.. Although it is not as universal as the traditional first-order SMC. 4.2 Aggregated HSMC 119 aggregated together to construct a higher sliding surface. According to the designed hierarchical structure. and the simulation results will be proposed. To construct the higher lay sliding surface. such a novel structure is entitled aggregated sliding surfaces. There is only a single-layer sliding surface.

However. To demonstrate the aggregated HSMC design.. 4. As a result. but also have to guar- antee that the sliding mode of the second-layer sliding surface is reachable. the aggregated sliding surfaces look like a simple combi- nation of the system state variables. s1 and s2. where the two subsystems are described by x_ 1 ¼ x2 ð4:5Þ x_ 2 ¼ f1 ðxÞ þ b1 ðxÞu .e. Figure 4. usw is the switching control law. In fact.120 4 Overhead Crane Control by Hierarchical Sliding Mode Fig. 4.. i. how to design the aggregated HSMC law and how to guarantee the stability analysis of each sliding surface.1. such a combination raises some novel control problems. Recall the crane model (2. How to design the equivalent-control-based aggregated HSMC remains problematic. α is treated as a constant here. the aggregated HSMC law is defined by u ¼ ueq1 þ ueq2 þ usw ð4:4Þ where ueq1 and ueq2 are the equivalent control laws of the two subsystems on their private first-layer sliding surfaces. or even a time-varying parameter. To attack the issue.16). the SMC law of the single-layer sliding surface is drawn from the reachability condition of SMC such that the SMC law can guarantee the system stability. there are two novel variables in the aggregated HSMC system. the hierarchical structure of the two-layer sliding surfaces has been constructed by capturing the physical nature of overhead cranes. the switching control law should not only make the closed-loop control system asymptotically stable. So far. the HSMC law has to stabilize the two variables by the design of the switching control law usw. i.1 illus- trates such a novel hierarchical structure of sliding surfaces.1 Hierarchical structure of the aggregated sliding surfaces Here α is constant and it can be a positive or negative parameter. the first-layer sliding surface variables. Concerning the control design.e. the equivalent-control-based SMC law on basis of the single lay sliding surface includes two parts: switching control law and equivalent control law. At the first glance. According to the hierarchical structure in Fig. Concerning the stability analysis.

adopt the aggregated HSMC law (4. Proof According to the Lyapunov’s stability scheme. design the hierarchical sliding surfaces (4.1 Consider the single-pendulum-type overhead crane system (2. substituting the subsystem model (4.2).3).6) into the derivatives of s2 yield c1 x2 þ f1 ðxÞ ueq1 ¼  ð4:7Þ b1 ðxÞ and c2 x4 þ f2 ðxÞ ueq2 ¼  ð4:8Þ b2 ðxÞ Compared with the aggregated HSMC law (4.2 Stability Analysis Theorem 4. differentiating s1 and s2 with respect to time t. The last term usw will be deduced from the system analysis to make the closed-loop control system and the two first-layer sliding surface variables asymptotically stable. ueq1 and ueq2 have been drawn in (4.4).2.5) into the derivatives of s1 and substituting the subsystems mode (4.4) is defined by ab1 ðxÞueq2 þ b2 ðxÞueq1 þ jS þ g sgnðSÞ usw ¼  ð4:9Þ ab1 ðxÞ þ b2 ðxÞ where κ and η are positive constants and α is a constant. a Lyapunov function can- didate can be defined by VðtÞ ¼ S2 =2 ð4:10Þ .7) and (4. the second-layer sliding surface S is of asymptotic stability if the switching control law in (4.16).4.1).7) and (4. From (4.8).8). (4. 4. To obtain the equivalent control laws of the two subsystems.1) and (4. Then.2).4).2 Aggregated HSMC 121 and x_ 3 ¼ x4 ð4:6Þ x_ 4 ¼ f2 ðxÞ þ b2 ðxÞu Adopt the equivalent-control-based SMC design. respectively. and define the equivalent control laws of the two systems (4. and (4. the equivalent control laws of the two subsystems will be applied to the two subsystems when the sliding modes of the two subsystems take place.

(4.7) and (4..2 Consider the single-pendulum-type overhead crane system (2. As a result.1). (4.14) to make the crane control system asymptotically stable.3) into the derivative of V yield _ VðtÞ ¼ Sða_s1 þ s_ 2 Þ ð4:11Þ Note that α is treated as a constant in (4.5) and (4. Theorem 4.12) yields _ VðtÞ ¼ S½aðc1 x2 þ f1 ðxÞ þ b1 ðxÞuÞ þ ðc2 x4 þ f2 ðxÞ þ b2 ðxÞuÞ ð4:13Þ Consider the aggregated HSMC law (4.9). and (4. the reachability condition of SMC is satisfied on the second-layer sliding surface and the sliding mode of the second-layer sliding surface is reachable in the finite time tf.13) yields _ VðtÞ ¼ Sfa½c1 x2 þ f ðxÞ1 þ b1 ðxÞ  ðueq1 þ ueq2 þ usw Þ þ ½c2 x4 þ f2 ðxÞ þ b2 ðxÞ  ðueq1 þ ueq2 þ usw Þg ð4:14Þ ¼ S½ab1 ðxÞ  ðueq2 þ usw Þ þ b2 ðxÞ  ðueq1 þ usw Þ _ From (4. .9) should be _ defined to achieve the condition VðtÞ\0. VðtÞ\0.6) into (4. VðtÞ\0 should exist in (4.1) and (4. _ Proof According to Theorem 4. and (4. the derivative of V can be written by _ VðtÞ ¼ S½aðc1 x_ 1 þ x_ 2 Þ þ ðc2 x_ 3 þ x_ 4 Þ ð4:12Þ Substituting the subsystem models (4. VðtÞ  0.2). Differentiating V with respect to time t in (4. Then.4).16). From (4.7).e.3). Substituting (4.8).14). and (4. the first-layer sliding surfaces s1 and s2 are asymptotically stable.4) and the equivalent control laws at the first-layer sliding surfaces (4.4). _ This fact indicates that SS\0. the second-layer sliding surface S is of asymptotic stability in the sense of Lyapunov.1. the motivation of the HSMC law is to first guarantee the stability the second-layer sliding surface.10) and VðtÞ\0 in (4. i. (4. From (4.2).122 4 Overhead Crane Control by Hierarchical Sliding Mode where S is the second-layer sliding surface variables. (4.7). consider the expressions of the two first-layer sliding surfaces (4. the aggre- gated HSMC law has the form of ab1 ðxÞueq1 þ b2 ðxÞueq2  jS  gsgnðSÞ u¼ ð4:15Þ ab1 ðxÞ þ b2 ðxÞ _ Since VðtÞ  0 in (4. Then.10). Further. adopt the aggregated HSMC law (4.11) and it can be treated as a time-varying one in the adaptive law design based on the aggregated hierarchical sliding surfaces. In the sense of Lyapunov.8).10) and substituting the expression of the second-layer sliding surface (4.10).8) into (4. design the hierarchical sliding surfaces (4.

there exists sitf ¼ si  ftf ðtÞ ð4:17Þ s_ itf ¼ s_ i  ftf ðtÞ ð4:18Þ where i = 1. on the subsequent time interval (tf. the system state trajectory keeps sliding on the second-layer sliding surface and converges to the origin as proven in Theorem 4. tf].18) mean the first-layer sliding surface variables s1 and s2 are bounded on the closed interval [0. Equations (4. define   S ¼ S 2 SjV_ ¼ 0 ð4:20Þ According to LaSalle’s principle. The fact indicates the phase trajectory of the second-layer sliding surface S will enter the neighborhood of the coordinate origin constructed by the axes s1 and s2 as t → ∞. VðtÞ   Sc ¼ S 2 R2 jV_  0 ð4:19Þ Further.21) Sc \ S ¼ Sc \ f as1 þ s2 ¼ 0 a_s1 þ s_ 2 ¼ 0 g ð4:22Þ As proven in Theorem 4. the system state trajectory moves toward the second-layer sliding surface S. the system trajectory tends to the largest invariant set inside S as t → ∞ in S \ Sc . To calculate the largest invariant set. On the time interval [0.14). To check the stabilities of s1 and s2 in [0. ∞). S \ Sc can be rewritten as   S c \ S ¼ Sc \ S ¼ 0 S_ ¼ 0 ð4:21Þ According to (4. On the sub- sequent open interval (tf.1. (4.1. ∞).3). tf] and both of them can converge to the second-layer sliding surface when the sliding mode is reached at tf. the time domain can be divided into two parts by the time point tf.17) and (4. define a truncated function as  1 t  tf ftf ðtÞ ¼ ð4:16Þ 0 t [ tf Consequently. the second-layer sliding surface variable S is asymptotically stable. Define a positively invariant and compact set From (4.2 Aggregated HSMC 123 Consequently. each subsystem states will move on the . the system trajectory maintains on the second-layer sliding surface S and the closed-loop crane control system becomes autonomous.22) can be drawn from (4. _  0. Subsequently. tf]. 2.4.

2. 4.3.05 Acceleration of gravity g (m s−2) 9. Such a small deviation is good enough in practice to increase crane effectiveness and throughout.2a–d. both the first-layer sliding surfaces s1 and s2 are of asymptotic stability. From Fig. the trolley velocity. and 4.1.3 Simulation Results Consider a single-pendulum-type overhead crane system. 4.1. the trolley can arrive at the origin from the initial position at about 6 s. the trolley accelerates as fast as possible to achieve the effective transport. From Fig. 4. 4.2.2. respectively. the largest invariant set inside S \ Sc is just the coordinate originate origin constructed by the sliding surface variables s1 and s2 rather than other points.2. where the trolley position. The parameters of the second-layer sliding surface is α = –2. obtained by a Simulink model of MATLAB in Appendix A. the maximum payload deviation is about 0.2b. and the payload angular velocity are displayed in Fig. 4. the payload angle.32 Payload mass m (kg) 5 Cable length L (m) 1.81 Initial state vector x0 [2 m 0 m s−1 0 rad 0 rad s−1] Desired state vector xd [0 m 0 m s−1 0 rad 0 rad s−1] .  s1 ¼ s_ 1 ¼ 0 ð4:23Þ s2 ¼ s_ 2 ¼ 0 As a result.14 rad. But it decelerates at about 1. Some numerical simulation results are demonstrated in Figs. Table 4.3. Some physical parameters of the crane [4] are determined in Table 4.4. 4.23) can be drawn. To verify the feasibility of the aggregated HSMC method for the crane system. some numerical simulation results are demonstrated. 4. The coefficients of the switching control law are determined by κ = 3 and η = 0. The parameters of the two first-layer sliding surfaces are selected as c1 = 0. 4. The crane performance is illustrated in Fig.8 s by the HSMC law to make the payload swing as small as possible. Displayed in Fig.7 and c2 = 8. In light of LaSalle’s principle.124 4 Overhead Crane Control by Hierarchical Sliding Mode first-layer sliding surface and toward its subsystem coordinate origin constructed by the subsystem state variables such that (4.2.2.1 Physical parameters and initial and desired states Trolley mass M (kg) 37. During the crane movement. the aggregated HSMC method can deal with the transport control problem of the crane system well and the HSMC system has a good performance.

4. a Trolley position. d Second-layer sliding surface S .3 Control performance of the aggregated SMC method. 4.2 System performance by the aggregated SMC method for the overhead crane. 4. b Trolley velocity. a Control input u. c Payload angle.2 Aggregated HSMC 125 Fig. c Subsystem sliding surface s2 at the first layer. d Payload angular velocity Fig. b Subsystem sliding surface s1 at the first layer.

126 4 Overhead Crane Control by Hierarchical Sliding Mode

The control performance is displayed in Fig. 4.3, where the control input, the
subsystem sliding surface s1 at the first layer, the subsystems sliding surface s2 at
the first layer, and the aggregated sliding surface S at the second layer are shown in
Fig. 4.3a–d, respectively.
From Fig. 4.3a, the control input has a chattering phenomenon. The aggregated
HSMC as a branch of SMC inherits the inner drawback of SMC as well. To
decrease the chattering, saturation function can be employed instead of signum
function in (4.15). In Fig. 4.3b–d, all the hierarchical sliding surfaces are stable as
proven in Theorems 4.1 and 4.2. In Fig. 4.3d, the sliding mode of the aggregated
sliding surface at the second layer is reachable at about 2 s. Thereafter, the two
sliding surfaces at the first layer tend to their subsystems origins along the two
subsystem sliding surfaces.
Although the two sliding surfaces at the first layer tend to be stable in Fig. 4.3b,
c, the reachability of their sliding modes cannot be theoretically guaranteed.
Theorem 4.2 can just guarantee that the two subsystem sliding surfaces are of
asymptotic stability. However, only the second-layer sliding surface S is reachable
in finite time because the reachability condition of sliding mode is guaranteed in
Theorem 4.1.
The phase curves of the aggregated sliding surfaces are shown in Fig. 4.4, where
the subsystem sliding surface s1 at the first layer, the subsystem sliding surface s2 at
the first layer, and the second-layer sliding surface S are displayed in Fig. 4.4a–c,
respectively.

Fig. 4.4 Phase portrait of the aggregated sliding surfaces. a Subsystem sliding surface s1 at the
first layer; b Subsystem sliding surface s2 at the first layer; c Second-layer sliding surface S

4.2 Aggregated HSMC 127

From Fig. 4.4a, b, the two sliding surface variables s1 and s2 are asymptotically
stable. In other words, the sliding modes of the two subsystem sliding surfaces are
reachable in infinite time. In Fig. 4.4c, the phase trajectory of the second-layer
sliding surface tends to the sliding surface. The trajectory remains on the sliding
surface once the control system enters the sliding mode. Thereafter, the two sub-
systems tend to their subsystem origins along the subsystem sliding surfaces s1 and
s2. The three phase curves demonstrate the technical contents in Theorems 4.1 and
4.2.
The theoretical analyses and numerical simulation results show the aggregated
HSMC method is a possible choice to achieve the transport control problem of
single-pendulum-type overhead crane systems. The main contributions of the
aggregated HSMC method are (1) the aggregated hierarchical structure of sliding
surfaces is constructed according to the crane physical property; (2) the sliding
mode of the second layer sliding surface is reachable in finite time; (3) the stability
of the two subsystem sliding surfaces at the first layer is guaranteed.

4.3 Incremental HSMC

Recall the crane structure in Fig. 2.1 and the state Eq. (2.16). There are four state
variables to describe the crane system. As a result, a novel type of sliding surface
structure [5] can be constructed by increasing the number of variables to produce
the hierarchical structure of sliding surfaces. The basic idea behind the type of
incremental hierarchical structure [6, 7] is as follows.
Consider the crane system (2.16). Select arbitrary two state variables to construct
the first-layer sliding surface. Then, increase one state variable to construct the
second-layer sliding surface with the 1st-layer sliding surface variable. This process
proceeds till all the system state variables are included. The incremental hierarchical
structure is developed by general increment of the system state variables. As the
crane model covers four state variables, the incremental sliding surfaces of the crane
control system have three layers, illustrated in Fig. 4.5.
Consider the aggregated hierarchical sliding surfaces in Sect. 4.2, which is
developed by capturing the crane’s physical property. Each sliding surface on the
first layer is designed for each subsystem. As a result, the aggregated HSMC can
only be employed for mechanical systems composed of several subsystems.
Compared with the aggregated sliding surfaces structure, the development of the
first-layer sliding surface is defined by arbitrary two state variables in the crane
model, which does not depend on the crane’s physical property. The incremental
hierarchical structure is achieved till all the system variables are covered by the
sliding surfaces. Thus, the incremental sliding surfaces structure can be applied to
mechanical systems that are not composed of subsystems.
Concerning the incremental sliding surfaces, two control problems rise . They
are similar to the aggregated HSMC. In the first-order SMC design, its control
design can guarantee the stability of the traditional single-layer sliding surface. The

128 4 Overhead Crane Control by Hierarchical Sliding Mode

Fig. 4.5 Hierarchical
structure of the incremental
sliding surfaces

incremental sliding surfaces in Fig. 4.5 contain three layers. How to design a
control law on basis of the incremental sliding surfaces is one crucial issue of such a
HSMC law.
The other control problem is about the stability of the incremental sliding sur-
faces. There are three layers sliding surfaces in the incremental SMC-based crane
control system. It is also eager to know how to theoretically guarantee the stability
of all the sliding surfaces. Besides, the incremental hierarchical structure is uni-
versal because it does not depend on the physical property of cranes. Consider a
system with n state variables. Adopt the incremental sliding surfaces. n − 1 layers
sliding surfaces exist. It is also a challenging problem to investigate the stability of
the n − 1 layers sliding surfaces of the system with n state variables.

4.3.1 Control Design

The first step of the incremental HSMC design is to construct the incremental
structure of sliding surfaces. Recall the crane model (2.16). There are four state
variables in the model. As mentioned, arbitrary two state variables can be employed
to formulate the first-layer sliding surface. Without loss of generality, the state
variables x1 and x2 in (2.16) are adopted such that the first-layer sliding surface has
the form

s 1 ¼ c1 x1 þ x2 ð4:24Þ

where c1 is a positive constant in order to have the stability of the first-layer surface.
The second layer sliding surface s2 is constructed by the linear combination of
the first-layer sliding surface variable s1 and one state variable. Without loss of
generality, x3 is picked up such that the second-layer sliding surface can be for-
mulated by

4.3 Incremental HSMC 129

s 2 ¼ c2 x3 þ s 1 ð4:25Þ

where c2 is constant.
In (4.24), x2 is equal to the derivative of x1. From the viewpoint of stability, c1
should be a positive constant. In (4.25), there is no explicit mathematic function
between x3 and s1. Here c2 is just a constant. Similarly, the third-layer sliding
surface can be written as

s 3 ¼ c3 x4 þ s 2 ð4:26Þ

where c2 is constant. So far, the incremental hierarchical structure of the sliding
surfaces has been built. The schematic is illustrated in Fig. 4.5.
Although the first-layer sliding surface is constructed by the two state variables
of the trolley subsystem, it is assumed that the two state variables are arbitrarily
picked up according to their footnote numbers. Compared with the aggregated
HSMC design, the three sliding surface variables in (4.24), (4.25), and (4.26) have
no explicit physical significance.
To achieve the transport control problem of the crane in Fig. 2.1, the incremental
HSMC law [3] is designed by

u ¼ ueq þ usw ð4:27Þ

In (4.27), ueq is the equivalent control law and usw is the switching control law.
Equation (4.27) has the same expression as the equivalent-control-based SMC law
on basis of the traditional single-layer sliding surface. Compared to the
equivalent-control-based first-order SMC design, Eq. (4.27) based on the incre-
mental sliding surfaces raises several novel problems, such as: How to design the
equivalent control law among the multiple-layer hierarchical sliding surfaces? How
to guarantee the stability of the multiple-layer hierarchical sliding surfaces?
To attack the issue of the equivalent control design in (4.27), the Lyapunov
function candidate is defined by

1
VðtÞ ¼ s23 ð4:28Þ
2

(4.28) means the control law (4.27) is deduced by making the last-layer sliding
surface asymptotically stable. Differentiating V in (4.28) with respect to time
t yields

dV dV ds3 d(c3 x4 þ s2 Þ
¼ ¼ s3 s_ 3 ¼ s3 ð4:29Þ
dt ds3 dt dt

Substituting (2.16), (4.24), (4.25), (4.26), and (4.27) into (4.29) yields

130 4 Overhead Crane Control by Hierarchical Sliding Mode

dV
¼ s3 ½c3 ðf2 þ b2 uÞ þ c2 x_ 3 þ s_ 1 
dt
¼ s3 ½c3 ðf2 þ b2 uÞ þ c2 x4 þ c1 x_ 1 þ x_ 2  ð4:30Þ
¼ s3 ½c3 ðf2 þ b2 uÞ þ c2 x4 þ c1 x2 þ f1 þ b1 u
¼ s3 ½c3 f2 þ c2 x4 þ c1 x2 þ f1 þ ðc3 b2 þ b1 Þðueq þ usw Þ

_
From (4.28), VðtÞ  0. In the sense of Lyapunov, VðtÞ\0 should exist in (4.30)
to have the stability of the last-layer sliding surface. To realize the purpose, define

c3 f 2 þ f 1 þ c2 x4 þ c1 x2
ueq ¼  ð4:31Þ
c 3 b2 þ b1

and

js3 þ gsgnðs3 Þ
usw ¼  ð4:32Þ
c 3 b2 þ b1

where κ and η are positive constants.
Substituting (4.31) and (4.32) into (4.30) yields

V_ ¼ js23  gjs3 j  0 ð4:33Þ

Since the derivative of V is just the left side of the reachability condition of the
last-layer sliding surface, (4.33) indicates that the last-layer sliding surface variable
s3 not only has the asymptotic stability, but also its sliding motion is reachable in
finite time.
So far, the first issue about the equivalent control design has been investigated.
But the stabilities of the other two inner sliding surfaces are kept to be unsolved
because the control law can only guarantee the stability of the last-layer sliding
surface rather than all the incremental sliding surfaces [8].
As far as the incremental HSMC design is concerned, the closed-loop crane
control system can be stabilized if and only if all the hierarchical sliding surfaces
are stabile. To have the stabilities of the other two inner sliding surfaces, more
insights and analyses about the incremental HSMC law (4.27) should be carried
out.

4.3.2 Stability Analysis

Theorem 4.3 Consider the single-pendulum-type overhead crane system (2.16),
design the incremental structure of the hierarchical sliding surfaces (4.24), (4.25),
and (4.26), define the control law (4.27), and adopt the equivalent control law

2) Stability of s2: Consider the third-layer sliding surface s3 achieves its sliding mode at tr. s3 would converge to a point on the phase plane by the axes x4 and s2 except the origin. s3 approaches the largest invariant set in   dV S ¼ s 3 2 Sc j ¼0 ð4:35Þ dt In (4. both the con- stants in (4.4. Moreover.e. which contradicts the fact limt!1 s3 ¼ 0. defined by   dV Sc ¼ s3 2 R j2 \c. the limit of s2 in (4.. the control law (4. all the incremental sliding surfaces are asymptotically stable if (4.37) must be equal to zero from the proof by contradiction. Consequently. According to LaSalle’s principle. there is no discontinuous switching control in the control system on the time interval (tr. ∞) and the closed-loop crane control system can be treated as an autonomous system. 3) Stability of s1: Since s2 is locally asymptotically stable.27) can guarantee the asymptotic stability of s3.34) is satisfied.32). the set Sc is a positively invariant and compact set.  c2 if x3 s1  0 c2 ¼ ð4:34Þ c2 if x3 s1 \0 Proof 1) Stability of s3: According to the control design. Then.25) can be drawn from lim s2 ¼ lim ðc2 x3 þ s1 Þ ¼ 0 ð4:38Þ t!1 t!1 Provided that the system states are not divergent. s3 js2 ¼ c3 x4 ¼ const: \ s_ 2 ¼ c3 x4 ¼ const:g Assume the two constants in (4.3 Incremental HSMC 131 (4. the limit of s1 can be deduced from . s2 is locally asymptotically stable. i.37) are not equal to zero as t → ∞. the last-layer sliding surface variable not only has the guaranteed stability. Thereafter. there exists S ¼ fs3 js3 ¼ 0 \ s_ 3 ¼ 0g ¼ fx4 . c[0 ð4:36Þ dt Thus. Then.35). but also its sliding mode is reachable at finite time tr. s3 jc3 x4 þ s2 ¼ 0 \ c3 x4 þ s_ 2 ¼ 0g ð4:37Þ ¼ fx4 .31) and the switching control law (4.

6a–d. c2 = –3. the first-layer sliding surface s1.3 Simulation Results To verify the feasibility of the incremental HSMC design for single-pendulum-type overhead crane systems. During the crane movement.2c. 4. The parameters of the three layers sliding surfaces are selected as c1 = 0. the velocity curve of the trolley in Fig.6. Figure 4. which indicates lim x3 ¼ 0 ð4:41Þ t!1 lim s1 ¼ 0 ð4:42Þ t!1 Equation (4.34) is satisfied.6. 4.6 can be obtained by a Simulink model of MATLAB in Appendix B.40) exists.3.7. where the trolley position. 4. and c3 = 0. Compared with Fig. Such a small deviation is good enough in practice to increase crane effectiveness and throughout.39) and (4. the incremental HSMC method can deal with the crane transport control problem and have a good performance.132 4 Overhead Crane Control by Hierarchical Sliding Mode lim s1 ¼  lim c2 x3 ¼ const: ð4:39Þ t!1 t!1 From (4. some numerical simulation results will be demonstrated. Moreover.85. For the purpose of comparison. The crane performance is illustrated in Fig. and the payload angular velocity are displayed in Fig.6c is slightly bigger. the physical parameters of the crane in Table 4. 4. the maximum payload deviation in Fig. where the control input. The coefficients of the switching control law are determined by κ = 3 and η = 0.40).34). the maximum payload deviation is about 0. 4.6. the reason is that the parameter c2 has to switch its sign according to (4. the payload angle. the trolley can arrive at the origin from the initial position at about 6 s. But both of the deviations are small enough in real applications. respectively.42) means the first-layer sliding surface variable s1 is locally asymp- totically stable if and only if the condition (4. The control performance is displayed in Fig. sgnðc2 x3 s1 Þ  0 ð4:40Þ According to (4. 4.6. 4.18 rad. the second-layer sliding surface s2. and the third-layer . From Fig.39) is only zero rather than other constants. As displayed in Fig. the trolley velocity.1 are adopted. the constant in (4.1. 4. Partly. (4.34) to keep the stability of the first-layer sliding surface s1.6b jumps back and forth. 4.4.2b is smooth but the trolley velocity in Fig. 4.

4. c Second-layer sliding surface s2. a Trolley position. b First-layer sliding surface s1. c Payload angle. a Control input u. d Payload angular velocity Fig. d Third-layer sliding surface s3 . 4.6 System performance by the incremental SMC method for the overhead crane. 4.3 Incremental HSMC 133 Fig.7 Control performance of the aggregated SMC method. b Trolley velocity.

9. the second-layer sliding surface s2.3.134 4 Overhead Crane Control by Hierarchical Sliding Mode sliding surface s3 are shown in Fig. The switch of c2 is shown in Fig.7a–d. 4. 4. and the Fig. 4. s2 tends to be stable and the switch of c2 is excited to make s1 asymptotically stable. respectively. a First-layer sliding surface s1. 4. b Second-layer sliding surface s2.9 Phase portrait of the incremental sliding surfaces.8. s3 enters its sliding mode at about 1 s. The phase curves of the incremental sliding surfaces are shown in Fig. where the first-layer sliding surface s1. the three layers sliding surfaces have the guaranteed stabilities.8 Switch process of the parameter c2 Fig. c Third-layer sliding surface s3 . 4. Thereafter. As proven in Theorem 4.

The advantage of such a combination is that the sliding mode of the sliding surface variable can be illustrated by the phase-plane method.3 Incremental HSMC 135 third-layer sliding surface s3 are displayed in Fig. From Fig.9a–c.9a.10 Hierarchical structure of the combing sliding surfaces . They can be divided into two groups. As suggested by its name. The derivative relations among the four state variables inspire us to define an intermediate variable z by combining x1 and x3 such that a sliding surface variable can be constructed by the intermediate variable and the derivative of the intermediate variable. 4.4 Combining HSMC Concerning the crane control problems. respectively. There are four state variables in (2. The schematic of the combining sliding surfaces is illustrated in Fig.16). However.9 jump back and forth when the sign of c2 changes. it is hard to depict the single-layer sliding surface of the first-order SMC method by phase portrait because of the constraint of the phase-plane method. the sliding surface variables of the first and second layers are just asymptotically stable. the sliding mode of s3 is reachable in finite time. Besides the aggregated and incremental HSMC methods. 4. 4. Fig. The aggregated and incremental HSMC methods over- come the drawback such that the phase portraits of the two HSMC methods can be intuitively illustrated in phase plane. 4. the com- bination of the state variables raises the stability problem that the four state vari- ables have no guaranteed stability. the combining HSMC design can be investigated. One group is composed of x1 and x3 and the other group covers x2 and x4.34) exists to guarantee the stability of the first-layer sliding surface. 10] can be constructed for the crane control problem.4. the combining hier- archical structure of the sliding surfaces [9. b. 4.9c. 4. the trajectories in Fig. because the sliding mode motion can only make the intermediate variable stable. Recall the mathematic model of the single-pendulum-type overhead crane (2. The basic idea behind the combining HSMC method [9. respectively. the challenge of the com- bining HSMC is how to combine the hierarchical structure of the sliding surfaces. where x2 and x4 are the derivatives of x1 and x3. 10] is as follows. On basis of the combining sliding surfaces. indicating they are reachable in infinite time.16). From Fig. Since the switch mechanism (4. 4.10.

1 Control Design Recall the crane model (2.16) into the derivative of s.45). Since the intermediate variable has the form of linear combination of x1 and x3. it is the source that the novel hierarchical structure of sliding surfaces is entitled the combining SMC. select the Lyapunov function candidate as VðtÞ ¼ s2 =2 ð4:47Þ Differentiating V in (4. Adopt the equivalent-control-based SMC design and define the combining SMC law as u ¼ ueq þ usw ð4:45Þ In (4. 4. 4.136 4 Overhead Crane Control by Hierarchical Sliding Mode 4.44) and the equivalent control law guarantees the system states keep sliding on the sliding surface and converge to the origin along the sliding surface. (4. and (4. differentiate s in (4. However. c is a positive constant.10.44) into the derivative of V yield . Define the intermediate variable z as z ¼ x1 þ cx3 ð4:43Þ In (4. cf2 ðxÞ þ f1 ðxÞ þ acx4 þ ax2 ueq ¼  ð4:46Þ cb2 ðxÞ þ b1 ðxÞ In order to have the stability of the top-layer sliding surface s.43).46) with respect to time t and substituting (2.46) with respect to time t and substitute the crane model (2. ueq is the equivalent control law and usw is the switching control law. which is determined by s ¼ az þ z_ ð4:44Þ here α should be positive to have the stability the top lay sliding surface. there is only the top lay sliding surface.10. where the switching control law is employed to drive the system states moving toward the specific sliding surface (4. the equivalent control law (4.4. the top lay sliding surface s is constructed by the intermediate variables and its derivative. To obtain the equivalent control law in (4.45). Then. Consider the group of state variables x1 and x3.46) can be deduced from s_ ¼ 0. such a structure is also treated as a hierarchical one because the intermediate variable and its derivative can be treated as the first-layer sliding surfaces. In Fig. From Fig.16).43).16).

The particularity of the combining SMC law is the x x intermediate variable results in a fixed ratio of 1 ¼ 2 .43). 4. the control law (4. which can facilitate the control design and analysis. it is necessary to draw the extra stability condition.45) in the sense of Lyapunov can just ensure that the intermediate variable z is asymptotically stable rather than the system states. indicating the stability of the top-layer sliding surface. adopt the control law (4.45). VðtÞ _ ¼ s_s\0 in (4. Assume that the sliding mode takes place at the time tf. construct the interme- diate variable (4.44).49). x3 x4 The combining SMC design employs the derivative relations between the state variables such that the sliding mode can be demonstrated in phase plane. the top lay sliding surface is of asymptotic stability. _ From (4.46) into (4.45) and the expression of the equivalent control law (4. In the sense of Lyapunov.4. substituting the control law (4. From the viewpoint of mathematics. this combining SMC is a subset of the traditional first-order SMC. In light of the reachability condition of sliding mode.2 Stability Analysis Theorem 4.47). However. then all the state variables are bounded. To asymptotically stabilize the whole system states. Further. VðtÞ  0.4 Combining HSMC 137 V_ ¼ s_s ¼ sða_z þ €zÞ ¼ sðax2 þ acx4 þ x_ 2 þ c_x4 Þ ð4:48Þ ¼ sfax2 þ acx4 þ f1 ðxÞ þ cf2 ðxÞ þ ½b1 ðxÞ þ cb2 ðxÞug Further.4.4 Consider the overhead crane system (2.49) ensures the sliding mode is reachable in finite time.50) into (4.16). design the combining sliding surfaces (4. .48) yield V_ ¼ sfax2 þ acx4 þ f1 ðxÞ þ cf2 ðxÞ þ ½b1 ðxÞ þ cb2 ðxÞðueq þ usw Þg ð4:49Þ ¼ s½b1 ðxÞ þ cb2 ðxÞusw Let usw ¼ js  gsgnðsÞ ð4:50Þ where κ and η are positive constants. VðtÞ\0 exists by substituting (4. the system trajectory moves along the sliding surface and converges to the coordinate origin constructed by the intermediate variable z and its derivative z_ . On the subsequent time interval.

By LaSalle’s principle. the limits of z and z_ are formulated by lim z ¼ lim ðx1 þ cx3 Þ ¼ 0 t!1 t!1 ð4:54Þ lim z_ ¼ lim ðx2 þ cx4 Þ ¼ 0 t!1 t!1 For arbitrary time t0 on the interval (tf. ∞).55) can be deduced from (4. This case indicates that the largest invariant set in E contains no sets other than the coordinate origin constructed by the intermediate variable and its derivative. there exist lim x1 ¼  lim cx3 ¼ const t!t0 t!t0 ð4:57Þ lim x2 ¼  lim cx4 ¼ const t!t0 t!t0 . the system trajectory converges to the coordinate origin constructed by the axes z and z_ along the surface s on the subsequent time interval (tf. the sliding surface s and the intermediate variable z are asymptotically stable as t → ∞ by means of Lasalle’s principle. s approaches the largest invariant set in   dV E ¼ sj ¼0 ð4:52Þ dt As mentioned. lim x1 ¼  lim cx3 t!t0 t!t0 ð4:55Þ lim x2 ¼  lim cx4 t!t0 t!t0 In (4. ∞). From (4. the intermediate variable z achieves the sliding mode at tf. Consequently.56) were satisfied.54). As a result. Thereafter.138 4 Overhead Crane Control by Hierarchical Sliding Mode Proof Define a set   X ¼ s 2 <2 jVðsÞ\e ð4:51Þ here ε is positive.56) contradicts (4.52).43) and (4.55). From (4. As ddVt \0. there would have existed lim x1 ¼  lim cx3 ¼ 1 lim x2 ¼  lim cx4 ¼ 1 ð4:56Þ t!t0 t!t0 t!t0 t!t0 Either of (4. Ω is a positively invariant and compact set. (4. if any state were divergent. the set E has the form E ¼ fsjs ¼ 0 \ s_ ¼ 0g ð4:53Þ ¼ fzjaz þ z_ ¼ 0 \ a_z þ €z ¼ 0g ¼ fzjz ¼ z_ ¼ €z ¼ 0g The set E is attracting.53).54) because the limit of z would be infinite as t → t0 if (4.

58) is satisfied on the time interval (tf. 3. x2. tf].62).. sup jxi j ¼ jjxi jj\1 ð4:59Þ t!tf here i = 1.43).e. it is apparent that Z1 Z1 2cx1 x3 dt\ ðx21 þ cx23 Þdt ð4:63Þ tf tf . all the state variables are bounded on [0..16).4 Combining HSMC 139 Compared with the asymptotically stable s and z. ∞) on account of the definition of the combining structure of the sliding surfaces. According to Theorem 4. i. 2. ∞) on account of the asymptotically stable intermediate variable z in [0. where tf is the time when the sliding mode of the sliding surface variable s takes place. ∞) into two parts.5 Consider the overhead crane system (2. construct the interme- diate variable (4.44). tf] and (tf. Equation (4. z 2 L2 on (tf.e. ∞). all the state variables are asymptotically stable if (4. design the combining sliding surfaces (4. Theorem 4. xj 2 L 1 and x_ j 2 L1 ð4:61Þ Further. 3. ∞) such that (4. [0.59) can be rewritten by sup jxj j ¼ jjxj jj\1 t!tf ð4:60Þ sup j_xj j ¼ jj_xj jj\1 t!tf here j = 1. Equation (4. i. all the state variables x1. ∞). x3. and x4 are just bounded on the time interval [0.4.45). 4.62) exists Z1 Z1 Z1 2 z dt ¼ 2 ðx1 þ cx3 Þ dt ¼ ðx21 þ cx23 þ2cx1 x3 Þdt\1 ð4:62Þ tf tf tf From (4.60). Then.61) can be drawn from (4. Subsequently. and adopt the control law (4.4.  c if x1 x3  0 c¼ ð4:58Þ c if x1 x3 \0 Proof The time tf divides the time interval [0. all the state variables are bounded on (tf. ∞).

63).81 Initial state vector x0 [0 m 0 m s−1 0 rad 0 rad s−1] Desired state vector xd [2.63) comes into existence as long as tf 4cx1 x3 dt [ 0. From (4.1. 4.2. some numerical simulation results will be demonstrated.65) can be deduced from (4.12.15 and are obtained by a Simulink model of MATLAB in Appendix C.58). Z1 Z1 Z1 2 4cx1 x3 dt\ ðx1 þ cx3 Þ dt ¼ z2 dt ð4:64Þ tf tf tf R1 Equation (4. which can be guaranteed by choosing the sign of c by means of (4. 4.140 4 Overhead Crane Control by Hierarchical Sliding Mode Further.487.0 m 0 m s−1 0 rad 0 rad s−1] .61) and (4. (4. The physical parameters of the overhead crane system and the initial and desired states [10] are shown in Table 4. 3.13.14. Table 4. Further.65). both x1 and x3 have the asymptotic stability in terms of Barbalat’s lemma on the time interval (tf. Some numerical results are shown in Figs.3 Simulation Results To verify the feasibility of the combining HSMC design for the single-pendulum-type overhead crane system.305 Acceleration of gravity g (m s−2) 9. ∞).65) means xj 2 L2 . x2 and x4 are of asymptotic stability as well because of the derivative relations among the state variables.11. The coefficients of the switching control law are determined by κ = 4 and η = 0. As a result.8 Cable length L (m) 0.2 Physical parameters and initial and desired states Trolley mass M (kg) 1 Payload mass m (kg) 0. 4. here j = 1. and 4. 4.4. Z1 Z1 x21 dt\1 and x23 dt\1 ð4:65Þ tf tf Equation (4. The parameters of the three layers sliding surfaces are selected as c = 0. 4.242 and α = 0.

4. b Intermediate variable z. a Trolley position. 4.11 System performance by the combining SMC method without parameter switch. b Trolley velocity. d Top-layer sliding surface s . a Control input u. c Derivative of the intermediate variable.12 Control performance of the combining SMC method without parameter switch.4 Combining HSMC 141 Fig. c Payload angle. d Payload angular velocity Fig.4.

b Intermediate variable z. b Trolley velocity.142 4 Overhead Crane Control by Hierarchical Sliding Mode Fig. a Control input u.14 Control performance of the combining SMC method with parameter switch. 4. d Payload angular velocity Fig.13 System performance by the combining SMC method with parameter switch. a Trolley position. c Payload angle. d Top-layer sliding surface s . 4. c Derivative of z.

11 and 4. The harmonic motions of the trolley and payload make the control input curve harmonic in Fig. .4 Combining HSMC 143 Fig. the combining SMC design arouses the novel problem of the crane control by SMC. the theoretical design cannot be put into practice because any real applications of crane systems cannot endure the harmonic motion of the trolley in Fig. and the top-layer sliding surface s.4. Such a linear combination of the state variables results in the bounded control system. 4. the payload angle. As proven in Theorem 4. The theoretical analysis coincides with the curves in Fig. 4. The asymptotic stability of the intermediate variable does not indicate the asymptotic stability of the state variables. 4. According to the combining HSMC design.11c also has the harmonic motion.4. where the trolley position. 4. 4. 4. all the state variables of the crane system are bounded. As displayed in Figs.12. respectively. the trolley velocity. 4. and the payload angular velocity are displayed in Fig.15 Switch process of the parameter c during the system dynamics The crane performance under the combining SMC design without the switch of c is illustrated in Fig. the intermediate variable. the sliding mode takes place at about 2. 4. Such swag is very small and it is acceptable in most real applications.11a.4 s. The control performance under the combining SMC design without the switch of c is illustrated in Fig.5. respectively. the control law can guarantee the asymptotic stability of the intermediate variable z. which can be treated as residual oscillations of the crane system.12. 4.11. 4.12a–d. From Fig.11a–d. 4. and the top-layer sliding surface are dis- played in Fig. The payload curve in Fig. where the control input.12d. This time point is very important for the switch of c in Theorem 4.12b–d. Although the crane control system is stable in the sense of Lyapunov. the derivative of the intermediate variable.12a.

the derivative of the intermediate variable. the sliding surfaces are fixed and they wait for the state trajectories entering the sliding mode. adaptive law design. The stabilities of the three kinds of crane control systems are also pre- sented. From Fig. i.5. and the payload angular velocity are displayed in Fig. Compared with the classic SMC designs in Chap.144 4 Overhead Crane Control by Hierarchical Sliding Mode The crane performance under the combining SMC design with the switch of c is illustrated in Fig. and so on. 4.e. Once the switch of c is carried out. The fact means that the crane per- formance by the combining HSMC law is not as effective as the other two HSMC methods. the switch of c enters the crane control system at 3 s. 4.14d. 4. As proven in Theorem 4.. respectively.3. some extensions can be explored. 3.14. The switch process of the parameter c is shown in Fig. all the state variables of the crane system can be asymptotically stable when the switch mechanism of the parameter c is carried out after the sliding mode of s is reached. the payload angle. To improve the crane control performance. the trolley velocity. respectively. Compared with the numerical results in Sects. which is slower than the aggregated and incremental HSMC methods. 4. where the control input. complex crane systems. Based the basic hierarchical structures. the three kinds of hierarchical structure of the sliding surfaces have been constructed for single-pendulum-type overhead cranes by employing the crane state variables. where the trolley position. 4. Some numerical results are illustrated. Thereafter.15.14a–d. one possible choice is to make fixed sliding surfaces active. which enrich the contents of SMC. So far.13.13 but rather get bounded in Fig.11. the intermediate variable.13a–d.5 Adaptive Control Design Based on Hierarchical Sliding Surfaces The aggregated. the crane control system by the combining HSMC law can arrive the desired position at about 10 s. the HSMC designs capture the physical nature of the cranes and present some novel control structures. The process continues till the crane control system arrives at the desired position. incremental and combining HSMC designs are basic types of hierarchical structures. Design of the adaptive sliding surfaces is just such a choice. 4. 4. these types of sliding surfaces are passive. Under any hierarchical structures. 4. the switch mechanism is triggered once the sign of x1 × x 3 changes. The control performance under the combining SMC design with the switch of c is illustrated in Fig. the state variables of the crane control system become asymptotically stable in Fig. 4. The adaptive sliding surfaces can move in the phase plane according to the . In this sense. In the designs. all the sliding surface parameters are con- sidered as time-invariant ones such that all the hierarchical sliding surfaces are linear. 4.2 and 4. and the top-layer sliding surface are dis- played in Fig.

Design the sliding surfaces of the trolley and payload subsystems as s 1 ¼ c1 x1 þ x2 ð4:66Þ s 2 ¼ c2 x3 þ x4 ð4:67Þ Here c1 and c2 are positive constants. Without loss of generality. the top-layer sliding surface S could be in any quadrant in the phase plane by the axes s1 and s2 since there are no differential relations between s1 and s2. The parameters c1 and c2 at the lower layer are fixed. To design the adaptive hierarchical sliding surfaces.5.4. According to the aggregated structure. Such an adaptive design can make the state trajectories may enter the sliding mode as soon as possible.66) and (4. adaptive control design based on hierarchical sliding sur- faces [11] is taken into consideration. On the other hand.16). To realize the purpose. respectively. only the parameter α at the top layer is considered as a time variable. the aggregated hierarchical sliding surfaces are employed for the adaptive law design. The motivations of such an idea are as follows.67).66) and (4. the adaptive law is considered for the purpose.67) are located at the first layer.5 Adaptive Control Design Based on Hierarchical Sliding Surfaces 145 designed adaptive law so that they become active to search the state trajectories. 4.1 Control Design Recall the crane model (2. Concerning the parameter α. The design method can be generalized to the incremental and combining sliding surfaces as well. x2 and x4 are the derivatives of x1 and x3 with respect to time t. the subsystem sliding surfaces s1 and s2 must be in the second and fourth quadrants to stabilize the sliding surface variables. aggregate the two sliding surfaces and design the second-layer sliding surface as S ¼ as1 þ s2 ð4:68Þ Here α is a time-varying parameter and it can be positive or negative.2 is designed at the top layer. Consequently. the two subsystems converge to these subsystem sliding surfaces by their subsystem equivalent control laws. respectively. Once the system trajectory enters the sliding mode at the top. To construct the higher layer sliding surface. The switching control law in Sect. To achieve the sliding mode at the top layer as soon as possible. . In (4. Only the other two subsystem sliding surfaces are not reachable in finite time. 4. both the subsystem sliding surfaces (4. only the top-layer sliding surface has the guaranteed reachability condition and it can enter the sliding mode in finite time.

5.146 4 Overhead Crane Control by Hierarchical Sliding Mode The aggregated HSMC law is also defined by u ¼ ueq1 þ ueq2 þ usw ð4:69Þ where the equivalent control laws of the two subsystems are formulated by 1 ueq1 ¼  ½c1 x2 þ f1 ðxÞ ð4:70Þ b1 ðxÞ and 1 ueq2 ¼  ½c2 x4 þ f2 ðxÞ ð4:71Þ b2 ðxÞ So far. a Lyapunov function candidate can be defined as 1 VðtÞ ¼ S2 ð4:74Þ 2 where S is the second-layer sliding surface variables. some parts of the adaptive control based on the aggregated sliding surfaces have been done. The adaptive law of α and the switching control law usw will be drawn from the system stability analysis. design the hierarchical sliding surfaces (4. (4. On basis of Lyapunov stability theory.66). the second-layer sliding surface S of the aggregated sliding surfaces is of asymptotic stability if the switching control law in (4.16).69). Then.6 Consider the single-pendulum-type overhead crane system (2. 4.2 Stability Analysis Theorem 4. Proof Similar to the proof of Theorem 4. and δ is a small positive constant to avoids s1 = 0.68). .1. and adopt the aggregated HSMC law (4.67). and (4.69) is defined by jS þ g sgnðSÞ usw ¼  ð4:72Þ ab1 ðxÞ þ b2 ðxÞ and the adaptive law of α is designed by b1 ðxÞueq2 s1 b1 ðxÞueq1 s1 a_ ¼ a  ð4:73Þ jjs1 jj2 þ d jjs1 jj2 þ d where κ and η are positive.

67) into (4.77) yields _ VðtÞ _ 1 þ a½c1 x2 þ f1 ðxÞ þ b1 ðxÞðueq1 þ ueq2 þ usw Þ ¼ Sfas þ c2 x4 þ f2 ðxÞ þ b2 ðxÞðueq1 þ ueq2 þ usw Þg ð4:78Þ _ 1 þ a½b1 ðxÞðueq2 þ usw Þ þ b2 ðxÞðueq1 þ usw Þg ¼ Sfas _ 1 þ ½ab1 ðxÞ þ b2 ðxÞusw þ ab1 ðxÞueq2 þ b2 ðxÞueq1 g ¼ Sfas _ From (4. design the hierarchical sliding surfaces (4. According to the control design of the adaptive control system.73).74).67).16).7 is similar to the proof of Theorem 4. Then. Proof The proof of Theorem 4. VðtÞ  0.3 Simulation Results To verify the feasibility of the adaptive control based on the hierarchical sliding surfaces.66) and (4. some numerical simulation results will be demonstrated. the 2nd-layer sliding surface S is of asymptotic stability in the sense of Lyapunov. To achieve this purpose.78).75) yields _ VðtÞ _ 1 þ aðc1 x_ 1 þ x_ 2 Þ þ c2 x_ 3 þ x_ 4  ¼ S½as ð4:76Þ Consider the crane model (2.69). the first-layer sliding surfaces s1 and s2 in (4.70) and (4.70). (4. VðtÞ\0 should exist in (4.4.5.5 Adaptive Control Design Based on Hierarchical Sliding Surfaces 147 Differentiating V with respect to time t in (4. 4.1. and (4.78) to make the crane control system asymptotically stable.68) into the derivative of V yield _ VðtÞ _ 1 þ s_ 2 Þ ¼ Sða_s1 þ as ð4:75Þ Substituting (4.69).74) and substituting the expression of the second-layer sliding surface (4. .16).66). define the switching control law as (4. adopt the aggregated HSMC law (4.70) and (4.71).72) and select the adaptive law of the _ parameter α as (4. and (4. Theorem 4.68). define the switching control law as (4.7 Consider the single-pendulum-type overhead crane system (2. define the equivalent control laws of the two sub- systems as (4. Equation (4.71) into (4. (4. Then VðtÞ\0 can be deduced from (4. The physical parameters of the overhead crane system and the initial and desired states are kept unchanged from Table 4.73).2. In the sense of Lyapunov.76) can be re-arranged as _ VðtÞ _ 1 þ a½c1 x2 þ f1 ðxÞ þ b1 ðxÞu þ c2 x4 þ f2 ðxÞ þ b2 ðxÞug ¼ Sfas ð4:77Þ Substituting (4.71) are asymptotically stable.72) and select the adaptive law of the parameter α as (4.

c Payload angle.16. But the adaptive law can apparently decrease the payload deviation.2.17. 4.16a–d. d Payload angular velocity The parameters of the two subsystem sliding surfaces at the first layer are selected as c1 = 0. 4. The crane performance under the adaptive control law based on the aggregated structure of sliding surfaces is illustrated in Fig. As illustrated in Fig. .16 System performance by adaptive control based on the aggregated HSMC. 4.16. where the trolley position. and the payload angular velocity are displayed in Fig.18.73) is set by 0. 4. 4. Some numerical results are displayed in Figs. 4. the trolley velocity.5.1. respectively. 4. the maximum payload deviation during the crane transport is about 0. and 4.7 and c2 = 8. 4. 4. the trolley can arrive the desired position at about 6 s. 4. the maximum angular deviation is just 0. 4. Such a small deviation is good enough in practice to increase crane effectiveness and throughput. the payload angle.12 rad. which are kept unchanged from Sect. which is 0. a Trolley position. In Fig.2.2c.02 rad smaller than the maximum deviation in Fig.16a.148 4 Overhead Crane Control by Hierarchical Sliding Mode Fig.16c. the designed adaptive law does not have the crane transport efficiency because the trolley also arrives the desired position at about 6 s in Fig.2. The coefficients of the switching control law are determined by κ = 3 and η = 0. Compared with the numerical results by the aggregated HSMC system in Fig.16a.12 rad. The initial value of α is determined by −2 and the value of δ in (4. b Trolley velocity.19 and are obtained by a Simulink model of MATLAB in Appendix D. Meanwhile. 4.

The aggregated sliding surface at the top layer in Fig.17a.18 Adaptive process of the parameter α versus time The control performance is displayed in Fig. the subsystem sliding surface s1 at the first layer.17a–d.6. where the control input.17 Control performance of adaptive control based on the aggregated HSMC. 4. which coincides with the results proven in Theorem 4. c Subsystem sliding surface s2. a Control input u.5 Adaptive Control Design Based on Hierarchical Sliding Surfaces 149 Fig. 4. respectively. 4. In Fig.17d is of asymptotic stability. d Sliding surface S Fig. 4. the subsystems sliding surface s2 at the first layer. b Subsystem sliding surface s1. 4. and the aggregated sliding surface S at the second layer are shown in Fig. On the other . the maximum control input is about 100 N.17. 4.4.

Tuan and Lee [12] investigated the transport control problem by the HSMC design for double-pendulum-type overhead cranes. 4. In Fig.18. 4.1 Control Design Recall the mathematical model of double-pendulum-type overhead cranes (2. the phase trajectory of the top-layer sliding surface is shown. From Fig. where the solid line is the phase trajectory and the dash line is the adaptive top-layer sliding surface.7. The type of cranes has three subsystems. 4.19. The adaptive process of the parameter α at the top-layer sliding surface is dis- played in Fig.6 HSMC Design for Double-Pendulum-Type Overhead Cranes The HSMC design can be extended to double-pendulum-type overhead cranes as well.150 4 Overhead Crane Control by Hierarchical Sliding Mode Fig.6.19 Phase portrait of the sliding surface S at the top layer hand. In .46).6 as t → ∞. indicating that the adaptive law can accelerate the phase trajectory to reach the sliding mode. the value of α is adaptively governed by the adaptive law designed in (4.17b.18. 4.73). the adaptive sliding surface can change its slope in the phase plane. 4. Adopting the aggregated structures of sliding surfaces. 4. Since such systems have one more subsystem compared to single-pendulum-type overhead cranes. but also challenges the system stability. c are asymptotically stable as proven in Theorem 4.46). 4. 4. From Fig.19. their control design not only becomes complex in mathematics. There are six state variables in (2. The value of α tends to be stabilized at about −3. the subsystem sliding surfaces in Fig.

The three subsystem sliding surfaces aggregated together to build up the first layer of the aggregated structure. the duties of ueq2 and ueq3 are to force the pairs of state variables (x3. design the sliding surface vector s to construct the aggregated structure of sliding surfaces and the vector s is described by 2 3 2 3 s1 k1 ðx1  xd Þ þ x2 s ¼ 4 s2 5 ¼ 4 k2 x 3 þ x 4 5 ð4:79Þ s3 k3 x 5 þ x 6 where λ1. respectively. To force state trajectories of the three subsystems to reach the first layer. x4) and (x5. let the derivative of s be zeroth.80). 2 3 2 3 2 3 s_ 1 k1 x_ 1 þ x_ 2 k1 x2 þ f1 ðxÞ þ b1 ðxÞueq1 s_ ¼ 4 s_ 2 5 ¼ 4 k2 x_ 3 þ x_ 4 5 ¼ 4 k2 x4 þ f2 ðxÞ þ b2 ðxÞueq2 5 ¼ 0 ð4:80Þ s_ 3 k3 x_ 5 þ x_ 6 k3 x6 þ f3 ðxÞ þ b3 ðxÞueq3 From (4. the equivalent control laws of the three subsystems are written as 1 ueq1 ¼  ½k1 x2 þ f1 ðxÞ b1 ðxÞ 1 ueq2 ¼ ½k2 x4 þ f2 ðxÞ ð4:81Þ b2 ðxÞ 1 ueq3 ¼ ½k3 x6 þ f3 ðxÞ b3 ðxÞ In (4. the equivalent control law ueq1 attracts a pair of state variables (x1. Each element in (4.79) with respect to time. Differentiate (4. a total equivalent control law is defined by ueq ¼ ueq1 þ ueq2 þ ueq3 ð4:82Þ In (4.80) can be obtained. to drive all state trajectories to reach the first layer of the aggregated structure of sliding surfaces. (4. [13] defined the equivalent control input determined from s_ ¼ 0.80). x6) to approach components s2 and s3. the switching control law usw must be introduced such that the HSMC law based on the aggregated sliding surfaces becomes u ¼ ueq þ usw ¼ ueq1 þ ueq2 þ ueq3 þ usw ð4:83Þ In (4.83). where S is referred as the aggregated sliding surface located at the top layer and the sliding surface is defined by .6 HSMC Design for Double-Pendulum-Type Overhead Cranes 151 [12]. x2) to component s1. Utkin et al. λ2.46) into the derivative of s.4. Then. and substitute (2. Similarly. To guarantee the reachability of sliding mode. all the parts are continuous. usw has a switching action with sufficiently high frequency to retain all state trajectories moving on manifold S = 0.82). and λ3 are constants. Therefore.79) depicts a subsystem sliding surface.

a Lyapunov function candidate can be defined by 1 V ¼ S2 ð4:85Þ 2 Differentiating V with respect to time t in time yields V_ ¼ SS_ ð4:86Þ Substituting the derivative of S into (4.83) into (4.152 4 Overhead Crane Control by Hierarchical Sliding Mode Fig. S denotes the switching line that all system states are attracted to and retained on.81) and (4.84). In (4. and μ3 are constants. To have the reachability of sliding mode at the aggregated sliding surface S. VðtÞ\0 should exist in (4.20. In the sense of Lyapunov. μ2. For this purpose. (4. VðtÞ  0. 4.46).84).20 Aggregated structure of sliding surfaces for double-pendulum-type overhead cranes S ¼ lT s ¼ l1 s1 þ l2 s2 þ l3 s3 ð4:84Þ where μ = [μ1 μ2 μ3]T and μ1. 4. From (4.86) yields V_ ¼ Sðl1 s_ 1 þ l2 s_ 2 þ l3 s_ 3 Þ ð4:87Þ Consider the double-pendulum-type overhead crane model (2. Then.85). and substitute (4. the switching control usw can be defined by .87) to make the crane control system asymptotically stable.87). the aggregated structure of sliding surfaces of the crane control systems can be demonstrated in Fig. V_ ¼ Sfl1 b1 ðxÞ½ueq2 þ ueq3 þ usw  þ l2 b2 ðxÞ½ueq1 þ ueq3 þ usw  þ l3 b3 ðxÞ½ueq1 þ ueq2 þ usw g ð4:88Þ ¼ Sf½l1 b1 ðxÞ þ l2 b2 ðxÞ þ l3 b3 ðxÞusw þ l1 b1 ðxÞðueq2 þ ueq3 Þ þ l2 b2 ðxÞðueq1 þ ueq3 Þ þ l3 b3 ðxÞðueq1 þ ueq2 Þg _ From (4.88) can be drawn.

design the hierarchical sliding surfaces (4.2 and μ3 = –0. The coefficients of the switching control law are determined by κ = 1 and η = 0.35. and define the switching control law as (4. s2 and s3 in (4.79) are asymptotically stable. the aggregated sliding surface located at the top layer is asymptotically stable.2 Stability Analysis Theorem 4.4.80) and (4. The stabilities of the three subsystem sliding surfaces have to be considered because they have no guaranteed stabilities from the control design. 4. The physical parameters of the overhead crane system are kept unchanged from Table 3. Proof The proof of Theorem 4.6.84).3 Simulation Results To verify the feasibility of the HSMC method for double-pendulum-type overhead crane systems. adopt the aggregated HSMC law (4. then the HSMC law based on the aggregated structure of sliding surfaces (4.35. respectively. μ2 = 1. 4.89). define the equivalent control laws of the three subsystems as (4.6 HSMC Design for Double-Pendulum-Type Overhead Cranes 153 l1 b1 ðxÞðueq2 þ ueq3 Þ þ l2 b2 ðxÞðueq1 þ ueq3 Þ þ l3 b3 ðxÞðueq1 þ ueq2 Þ usw ¼  l1 b1 ðxÞ þ l2 b2 ðxÞ þ l3 b3 ðxÞ ð4:89Þ jS þ g sgnðSÞ  l1 b1 ðxÞ þ l2 b2 ðxÞ þ l3 b3 ðxÞ where κ and η are positive constants. Some numerical .46). c2 = 22 and c3 = 50.2. the subsystem sliding surfaces located at the 1st-layer s1.02. 2 and the initial and desired states are set by x0 = [0 m 0 m s−1 0 rad 0 rad s−1 0 rad 0 rad s−1] and xd = [4 m 0 m s−1 0 rad 0 rad s−1 0 rad 0 rad s−1].6.81) and (4.8 is very similar to the proof of Theorem 4.83). Substitute (4.90) has the form l1 b1 ðxÞueq1 þ l2 b2 ðxÞueq2 þ l3 b3 ðxÞueq3 u¼ l1 b1 ðxÞ þ l2 b2 ðxÞ þ l3 b3 ðxÞ ð4:90Þ jS þ g sgnðSÞ  l1 b1 ðxÞ þ l2 b2 ðxÞ þ l3 b3 ðxÞ From the control design. The parameters of the three subsystem sliding surfaces at the first layer are selected as c1 = 0.8 Consider the double-pendulum-type overhead crane system (2. some numerical simulation results will be demonstrated.83).89) into (4. Then.18. 81). The parameters of the aggregated sliding surface at the top layer are selected as μ1 = 1.

22. where the trolley position. the subsystem sliding surface s1. 4. the subsystems . 4.22 and are obtained by a Simulink model of MATLAB in Appendix E. However. a Trolley position. From Fig.154 4 Overhead Crane Control by Hierarchical Sliding Mode Fig. residual oscillations of the double-pendulum dynamics are hard to be completely suppressed by the HSMC method in practice. Meanwhile.21. respectively. the subsystems sliding surface s2. where the control input. the payload angle. The control performance is displayed in Fig. the trolley velocity. As a result. the trolley arrives the desired position at about 12 s. the hook angle. and the payload angular velocity are displayed in Fig. On the other hand. the hook angular velocity. 4.21a–f. e Payload angle.21 System performance by the HSMC law for double-pendulum-type overhead cranes. 4.1 rad.21. d Hook angular velocity. Such deviations are small enough in real crane applications. 4.21. b Trolley velocity. both the maximum angular deviations of the hook and payload are not larger than 0. The crane performance under the HSMC law is illustrated in Fig. f Payload angular velocity results are displayed in Figs. 4. such a perfect model is too ideal to obtain. and 4. the performance of the two angular curves by the HSMC method has no residual oscillations under the condition that the crane model has no modeling errors. c Hook angle.

4. which is an inner drawback of the SMC method. the control input has chattering. only the reachability condition of the sliding surface S at the top layer is guaranteed during the control design. Further.22e. 4. The whole sliding surfaces in Fig. d Subsystem sliding surface s3. concerning these sliding surfaces of the aggregated hierarchical struc- ture. 4. From Fig. respectively.8. and the aggregated sliding surface S are shown in Fig.6 HSMC Design for Double-Pendulum-Type Overhead Cranes 155 Fig. c Subsystem sliding surface s2. the aggregated sliding surface at the top layer enters its sliding mode at about 3 s.22a.4.22a–e. Thereafter. s2. the trajectories of the three subsystems in phase plane enter their private subsystem sliding surfaces under the actions of their private subsystem equivalent controls. The three subsystem sliding surfaces s1. b Subsystem sliding surface s1. e Top-layer sliding surface S sliding surface s3. and s3 are just asymptotically stable rather than being reachable in finite time. . From Fig. 4. 4. a Control input u.22 Control performance by the HSMC law for double-pendulum-type overhead cranes.22b–e are asymptotically stable as proven in Theorem 4.

Chap.x.num2str(flag)]). 3. [sys. sizes. Controller program: AHSMC.u.str = []. 4. case {1. m. sizes. sizes.NumOutputs = 4. sizes.str.ts]=mdlInitializeSizes sizes = simsizes.x0.ts] = sfuntmpl(t.4. The file is very similar to the SPCrane.DirFeedthrough = 1.156 4 Overhead Crane Control by Hierarchical Sliding Mode Appendices A Simulink Model to Plot Figs. case 0. x0 = []. end function [sys.x0.NumContStates = 0. otherwise error(['Unhandled flag = '.str. % at least one sample time is needed sys = simsizes(sizes).flag) switch flag. .ts]=mdlInitializeSizes.2.NumDiscStates = 0.4 Plant program: SPCrane.3 and 4. m function [sys.u).NumSampleTimes = 0.x. case 3.9} sys=[].x0.ts = [].str. sizes.2. 4. sizes.NumInputs = 4. sys=mdlOutputs(t.m file in Appendix A.

bata=[1 1]. 4. b1 =D*inv(A*D-B*C).l=1. s2=afa(2)*e(3)+bata(2)*e(4). uslaw=-num*inv(den)+ds*inv(den).7 8.C=(cos(u(3))). s1=afa(1)*e(1)+bata(1)*e(2).Appendices 157 function sys=mdlOutputs(t.7. k=3.x. % parameter of controller e(1)=(u(1)). a =-2. ds=-k*s-ita*sign(s).D=l. sys(2)=s1.S2=a*S1+s2.1. sys(3)=s2.32.e(2)=u(2).e(3)=u(3). num=a*(afa(1)*u(2)+bata(1)*f1)+(afa(2)*u(4)+bata(2)*f2). 4.e(4)=u(4).B=(m*l*cos(u(3))). B Simulink Model to Plot Figs. 4.s=S2.6. f2 =(m*l*u(4)^2*sin(u(3))*C+g*A*sin(u(3)))*inv(B*C-A*D).3. b2 =C*inv(B*C-A*D).M=37.05. sys(1)=uslaw. den=a*bata(1)*b1+bata(2)*b2.9 .u) % parameter of system g=9.8 and 4. sys(4)=S2.2].m=5. f1 =(m*l*u(4)^2*sin(u(3))*D+g*B*sin(u(3)))*inv(A*D-B*C).81.S1=s1.E=(m*l*sin(u(3))).afa =[0.ita=0. A=(M+m).

NumDiscStates = 0. sizes. [sys.4].x.} sys=[].85 1 -3. m function [sys. case {1.u) % system prameters g=9.NumContStates = 0. c=[0.ts]=mdlInitializeSizes.NumSampleTimes = 0.6 0. case 3.x0.158 4 Overhead Crane Control by Hierarchical Sliding Mode Plant program: SPCrane.m file in Appendix A.NumInputs = 4. . Controller program: IHSMC.. Chap. k=3.ita=0. sizes.x0.. (m*l*(cos(u(3)))^2-(M+m)*l). 3. b1=1/(M+m-m*(cos(u(3)))^2). m. sizes. s2=s1+c(3)*u(3). The file is very similar to the SPCrane.num2str(flag)]).M=37.ts]=mdlInitializeSizes sizes = simsizes. f2=(m*l*u(4)^2*sin(u(3))*cos(u(3))+(M+m)*g*sin(u(3)))/.str.NumOutputs = 5. x0 = [].2.1. sizes. s=s2+c(4)*u(4). % at least one sample time is needed sys = simsizes(sizes).ts = []. f1=(m*l*u(4)^2*sin(u(3))+m*g*sin(u(3))*cos(u(3)))/(M+m-m*(cos(u(3)))^2).x. s1=c(1)*u(1)+c(2)*u(2).u). otherwise error(['Unhandled flag = '.05.m=5.9.x. function sys=mdlOutputs(t. case 0.l=1.81. sys=mdlOutputs(t.x0.str = [].u. sizes.ts] = sfuntmpl(t.str. sizes.DirFeedthrough = 1.flag) switch flag. b2=(cos(u(3)))/(m*l*(cos(u(3)))^2-(M+m)*l).str. end function [sys.4.32.

11.15 . sys(1) = uslaw. C Simulink Model to Plot Figs.14 and 4. 4. uslaw=-num*inv(den)+ds*inv(den). sys(5)=c(3).12. sys(3)=s2. den=c(2)*b1+c(4)*b2. sys(2) =s1.Appendices 159 if u(3)*s1<0 c(3)=-c(3). end ds=-k*s-ita*sign(s). sys(4)=s. 4. 4.num=c(1)*u(2)+c(2)*f1+c(3)*u(4)+c(4)*f2.13. 4. else c(3)=c(3).

305. [sys.8.k0=8.c2=c(2).4. str = [].num2str(flag)]).x.str.04. % at least one sample time is needed sys = simsizes(sizes).u) % system prameters M=1. function sys=mdlOutputs(t. .NumOutputs = 5.81.L=0.x. sizes.x0. f1=(m*L*u(4)^2*sin(u(3))+m*g*sin(u(3))… *cos(u(3)))*inv(M+m*sin(u(3))*sin(u(3))). sizes.x. sizes.} sys=[]. 3.m file in Appendix A. sizes. b2=-cos(u(3))*inv((M+m*sin(u(3))*sin(u(3)))*L).2. Controller program: CbHSMC. ts = [].u).160 4 Overhead Crane Control by Hierarchical Sliding Mode Plant program: SPCrane.NumInputs = 4. The file is very similar to the SPCrane.str.eta=0.flag) switch flag.DirFeedthrough = 1. x0 = [].ts] = sfuntmpl(t. case 0.m=0.x0. sizes. b1=1*inv(M+m*sin(u(3))*sin(u(3))).487].242 0.x0. f2=-((M+m)*g*sin(u(3))+m*L*u(4)^2*sin(u(3))*cos(u(3))) ……*inv((M+m*sin(u(3))*sin(u(3)))*L).g=9. sys=mdlOutputs(t. m.9.NumDiscStates = 0.str. sizes.ts]=mdlInitializeSizes sizes = simsizes. case 3.ts]=mdlInitializeSizes. c = [0. end function [sys. m function [sys.u.NumSampleTimes = 0. c1=c(1). case {1. Chap. otherwise error(['Unhandled flag = '.NumContStates = 0.

uslaw=-num*inv(den)+ds. D Simulink Model to Plot Figs. % end % end z=(u(1))+c1*u(3). sys(4) =s.17.16. 4. den=c1*b2+b1. ds=-k0*s-eta*sign(s). sys(5) =c1. sys(3) =dz. s=c2*z+dz. dz=u(2)+c1*u(4).18 and 4.4 % if u(1)*u(2)>=0 % c1=c1. 4. sys(1) = uslaw. % else % c1=-c1. 4.19 . num=(c1*f2+f1+c2*u(2)+c2*c1*u(4)). sys(2) =z.Appendices 161 % Stabilize all the state variables % if t>2.

E=(m*l*sin(u(3))). b2 =C*inv(B*C-A*D).ts]=mdlInitializeSizes sizes = simsizes. 3.num2str(flag)]).x. case 0.6.4. f1 =(m*l*u(4)^2*sin(u(3))*D+g*B*sin(u(3)))*inv(A*D-B*C). Controller program: AdaptHSMC.x0.NumOutputs = 5. sys=mdlDerivatives(t.x0.str. x0 = [-2]. m function [sys. The file is very similar to the SPCrane. b1 =D*inv(A*D-B*C).NumContStates = 1.str.u). afa =[0. sys=mdlOutputs(t.ts]=mdlInitializeSizes. case 3. function sys=mdlDerivatives(t.B=(m*l*cos(u(3))). [sys.81.NumDiscStates = 0.DirFeedthrough = 1.str. f2 =(m*l*u(4)^2*sin(u(3))*C+g*A*sin(u(3)))*inv(B*C-A*D).m=5. delta=0. case {2.C=(cos(u(3))). s1=afa(1)*u(1)+u(2). Chap. case 1.D=l.7 8.05. ueq1=-(afa(1)*u(2)+f1)*inv(b1).ts] = sfuntmpl(t. ts = [].u) % parameter g=9. ueq2=-(afa(2)*u(4)+f2)*inv(b2). sizes. str = [].2].u. otherwise error(['Unhandled flag = '.32. sys(1)=-x(1)*b1*ueq2*s1*inv(s1^2+delta)-(b2)*ueq1*s1*inv(s1^2+delta). m.flag) switch flag.x0. . A=(M+m). % at least one sample time is needed sys = simsizes(sizes).x.M=37.162 4 Overhead Crane Control by Hierarchical Sliding Mode Plant program: SPCrane.NumInputs = 4. end function [sys.NumSampleTimes = 0.x. sizes.l=1. sizes.m file in Appendix A.9} sys=[]. sizes.x. sizes. sizes.u).

M=37.Appendices 163 function sys=mdlOutputs(t.e(2)=u(2). S1=s1. s1=afa(1)*e(1)+bata(1)*e(2). e(1)=(u(1)). s2=afa(2)*e(3)+bata(2)*e(4). sys(3)=s2.05. f2 =(m*l*u(4)^2*sin(u(3))*C+g*A*sin(u(3)))*inv(B*C-A*D).l=1. f1 =(m*l*u(4)^2*sin(u(3))*D+g*B*sin(u(3)))*inv(A*D-B*C).e(4)=u(4). sys(2)=s1. num=x(1)*(afa(1)*u(2)+bata(1)*f1)+(afa(2)*u(4)+bata(2)*f2).1.C=(cos(u(3))). den=x(1)*bata(1)*b1+bata(2)*b2.D=l.B=(m*l*cos(u(3))). s=S2. ds=-k*s-ita*sign(s). afa =[0. S2=x(1)*S1+s2.7 8. sys(1)=uslaw.E=(m*l*sin(u(3))).m=5. sys(4)=S2.x. .81. k=3. uslaw=-num*inv(den)+ds*inv(den). b1 =D*inv(A*D-B*C). sys(5)=x(1).u) % parameter of system g=9. bata=[1 1].e(3)=u(3).32.2]. ita=0. b2 =C*inv(B*C-A*D). A=(M+m).

theta=0.x. otherwise error(['Unhandled flag = '.u.21 and 4.str = []. m function [sys. m. %desired parameter dis=4.ts] = sfuntmpl(t. sizes.str.DirFeedthrough = 1.l2=0. sizes.a2=22.164 4 Overhead Crane Control by Hierarchical Sliding Mode E Simulink Model to Plot Figs.ts = []. sys=mdlOutputs(t.l1=3.ddis=0. case 0.4. A=(mt+mh+mc).35. case 3. sizes.a1=0.b1=1.x.3.B=(mh+mc).22 Plant program: DPCrane.x0. s3=(u(7)-dtheta)+a3*(u(6)-theta). sizes.m file in Appendix B.x0.18.dphi=0.mh=10. sizes.02.u) %system parameter mt=50. %first_level sliding surface s1=(u(2)-ddis)+a1*(u(1)-dis).g=9. .x0.9} sys=[].35. s2=(u(4)-dphi)+a2*(u(3)-phi).NumContStates = 0.ante=0.a3=50.phi=0. function sys=mdlOutputs(t. [sys.NumOutputs = 5.b=0.u). 3.NumSampleTimes = 0.81.num2str(flag)]).b3=-0.b2=1.ts]=mdlInitializeSizes. end function [sys.NumDiscStates = 0. The file is very similar to the DPCrane. case {1.x. %designed parameter K=1.2. sizes.str.NumInputs = 8. % at least one sample time is needed sys = simsizes(sizes).flag) switch flag.str. Chap.mc=2. x0 = []. 4.2.dtheta=0.ts]=mdlInitializeSizes sizes = simsizes. Controller program: HSMC. %second_level sliding surface S=b1*s1+b2*s2+b3*s3.

UK. Italy. China.. sys(5)=S. gq=b1*(B*l1*cos(u(3))*u(5)+mc*l2*cos(u(6))*u(8)-b*u(2)- …. Wang W.. sys(3)=s2. Saridis GN (1983) Intelligent robotic control. Wang W. pp 9320–9325 10. sys(1)=control_law. Yi J. (l2*(mc*(cos(u(6)))^2-A))+b1*a1*(u(2))+b2*a2*(u(4))+b3*a3*(u(7)). Beijing. l1*B^2*cos(u(3))*sin(u(3))*(u(4))^2… +B*cos(u(3))*b*u(2))/(l1*B*(B*(cos(u(3)))^2-A)). +b2*(mc*l2*(mt*cos(u(3)-… u(6))+B*sin(u(3))*sin(u(6)))*u(8)+B*A*g*sin(u(3)). information and control. IEEE Trans Autom Control 28(5):547–557 2. Yi JQ (2013) Design of combining sliding mode controller for overhead crane systems. +mc*l2*(mt*sin(u(3)-u(6))+B*sin(u(3))*cos(u(6)))*(u(7))^2-.. Qian DW. Yi JQ. control_law=-inv(fq)*(gq+K*S+ante*sign(S)). Sanya. Qian DW (2007) Incremental sliding mode controller for large-scale underactuated system. Zhao D. References 1..B*l1*sin(u(3))*(u(4))^2-mc*l2*sin(u(6))*(u(7))^2)/A. Xi’an. Yi JQ (2010) Fuzzy aggregated hierarchical sliding mode control for underactuated systems... Yi JQ. pp 276–280 9. In: Proceedings of the 18th IFAC World Congress. Yi JQ. sys(2)=s1.. In: Proceedings of international conference on mechatronics and automation.. Wang W (2006) Proposal of incremental sliding mode control. Chengdu. Zhao DB (2011) Control of overhead crane systems by combining sliding mode with fuzzy regulator. China. sys(4)=s3. +b3*(l1*(mt*cos(u(3)-u(6))+B*sin(u(3))*sin(u(6)))*u(5)… +cos(u(6))*b*u(2)+A*g*sin(u(6)). Liu XJ.Appendices 165 %control law fq=b1/A-b2*cos(u(3))/(l1*(B*(cos(u(3)))^2-A))- …b3*cos(u(6))/(l2*(mc*(cos(u(6)))^2-A)). +l1*(B*sin(u(6))*cos(u(3))-mt*sin(u(3)-… u(6)))*(u(4))^2+mc*l2*cos(u(6))*sin(u(6))*(u(7))^2)/. pp 196–201 5. Milano. Zhao DB. Int J Control Autom 6(1):131–140 11. In: Proceedings of international conference on artificial intelligence and computational intelligence. In: Proceedings of IEEE international conference on networking. Qian DW. IEE Proc Control Theory Appl 151(6):683–690 3.. Hao YX. pp 1050–1055 . In: Proceedings of 2010 international conference on mechatronics and automation. IEEE Trans Autom Control 47(10):1696–1700 6. pp 340–343 7. China. Hao YX. Zhao DB. Qian DW.. Qian DW. Ma YF (2010) Fuzzy incremental hierarchical sliding mode control for underactuated systems. Liu D (2004) Design of a stable sliding-mode controller for a class of second-order underactuated systems. sensing and control. In: Proceedings of first international conference on innovative computing.. London. IET Control Theory Appl 1(1):163–172 4. Yi JQ (2007) Structure design of two types of sliding-mode controllers for a class of under-actuated mechanical systems. Yi JQ.. Qian DW. Liu XD. China. Yi JQ (2012) Adaptive control based on hierarchical sliding mode for under-actuated systems. Hirschorn RM (2002) Incremental sliding mode control of the ball and beam. pp 87–92 8.

Taylor & Francis.166 4 Overhead Crane Control by Hierarchical Sliding Mode 12. USA . J Mech Sci Technol 27(6):1863–1873 13. 2nd edn. Guldner J. Utkin V. Lee SG (2013) Sliding mode controls of double-pendulum crane systems. Tuan LA. Shi J (2009) Sliding mode control in electromechanical systems.

20) cover both matched and unmatched uncertainties. Without loss of generality. the chapter proceeds with source codes of all the simulations. ξ1 and ξ2 in (2. Keywords Overhead crane control  Uncertainty  Compensator design  Fuzzy  logic Neural networks 5..20) and (2. Qian and J.e. Some design methods will be developed to achieve the robustness of crane control based on SMC. and compensator design based on neural networks (NNs). These design methods can be extended to double-pendulum-type overhead cranes as well. As far as crane control is concerned.20).1007/978-3-662-48417-3_5 . the model of single-pendulum-type overhead cranes is employed to illustrate the design methods of compensator. Due to the invariance of sliding mode control (SMC).47). Whatever types of overhead crane systems are. Hierarchical Sliding Mode Control for Under-actuated Cranes. a crucial issue is how to deal with unmatched uncertainties. an overhead crane is adopted as benchmark. compensator design based on hierarchical sliding surfaces. i. This chapter attacks the issue. but overhead crane models cover uncertain terms. compensator design based on fuzzy logic. The uncertainties cover both matched and unmatched ones. In addition.1 Problem Description Recall uncertain crane models (2. unmatched uncertainties become challenging. For the purpose of illustration. The methodology of sliding mode control (SMC) can sup- press matched uncertainties in light of the invariance of sliding mode. © Springer-Verlag Berlin Heidelberg 2016 167 D. Three design methods are taken into account. and each control method is carried out by the transport control problem of the crane. matched parts of uncertainties in the crane model can be overcome. This chapter focuses on compensator design based on sliding mode for uncertain overhead cranes. DOI 10.Chapter 5 Compensator Design Based on Sliding Mode for Uncertain Overhead Cranes Abstract Crane systems have uncertainties. Concerning the crane control problem. Consider the crane mode (2. Yi.

Thus. Consider the nominal crane model (2. and its expression will be deduced from the stability analysis of the crane control system. ucn is kept unknown. To have the stability of the sliding surface S. 4. some expressions about the control law un for the nominal crane system will be directly employed without more explanations.3). where the unmatched uncertainties cannot be suppressed by the invariance of SMC. a lumped sliding mode compensator at the top layer is designed in Sect.1).1 Control Design Generally speaking. un is the aggregated HSMC law and its design and analysis have been introduced in detail in Sect.2. 4. Design the aggregated hierarchical sliding surfaces illustrated in Fig.e. the control input u of the uncertain crane system is defined by u ¼ un þ ucn ð5:1Þ where ucn is the law of compensation based on the aggregated hierarchical sliding surfaces.16). In (5. a Lyapunov function candidate is defined by . Model uncertainties are inevitable in practice.1 and formulate the sliding surfaces as (4.. the stability of the top-layer sliding surface S should be first considered. there are two methods to design a compensator on basis of the hierarchical sliding surfaces [1]. including parameter fluctuations.2.2. 5. The other method is to design a lumped compensator and compensate the unmatched uncertainties at the top layer. One is to design a distributed compensator and compensate the unmatched uncertainties at every layer of the hierarchical sliding mode surfaces [2].2) and (4. it will affect the stabilities of higher layers.1). Consider the crane mode (2. cover both matched and unmatched uncertainties. However. The uncertain crane mode (2. Many design methods based on SMC have been introduced in Chaps. 3 and 4. model uncertainties and random external disturbances. Its advantage is that this method simplifies the control design. The uncertainties. Section 5.20). The crane dynamics are subject to uncertainties. Because the lumped compensator is located at the top layer of the hierarchical sliding surfaces.2 investigates the problem of compensator design based on the aggregated hierar- chical SMC. To overcome the uncertainties.20) can be divided into two parts. nominal crane model and uncertainties.2 Compensator Design Based on HSMC Crane systems are inherently uncertain and nonlinear. Two disadvantages of this idea are that this makes the controller structure complex and that if the compensator at a lower layer does not eliminate the uncertainties. 5. (4.168 5 Compensator Design Based on Sliding Mode … 5. i.

5) yields _ VðtÞ ¼ SfjS  g sgnðSÞ þ ½ab1 ðxÞ þ b2 ðxÞucn þ an1 þ n2 g ð5:6Þ ¼ jS2  gjSj þ Sf½ab1 ðxÞ þ b2 ðxÞucn þ an1 þ n2 g Define the compensation law based on the hierarchical sliding surfaces as d0 sgnðSÞ ucn ¼  ð5:7Þ ab1 ðxÞ þ b2 ðxÞ where d0 [ 0 and d0  sup jan1 j þ sup jn2 j.8) because of d0  sup jan1 j þ sup jn2 j. Substitute (5.20). By means of the control scheme. As a result. Substituting (4. Then.4). the HSMC law of the nominal crane system in (4.3) yields _ VðtÞ ¼ S½aðc1 x_ 1 þ x_ 2 Þ þ c2 x_ 3 þ x_ 4  ¼ S½ac1 x2 þ af1 ðxÞ þ ab1 ðxÞu ð5:4Þ þ c2 x4 þ f2 ðxÞ þ b2 ðxÞu þ an1 þ n2  Substitute the control input (5. Then.7) into (5.1) into (5. the aggregated sliding surface S located at the top layer is of asymptotic stability in the crane control system with unmatched uncertainties.5.3) can be obtained.2 Compensator Design Based on HSMC 169 S2 VðtÞ ¼ ð5:2Þ 2 Differentiate V with respect to time t in (5. _ VðtÞ ¼ SS_ ¼ Sða_s1 þ s_ 2 Þ ð5:3Þ Consider the crane mode (2.4) becomes _ VðtÞ ¼ Sfac1 x2 þ af1 ðxÞ þ ab1 ðxÞun þ c2 x4 þ f2 ðxÞ þ b2 ðxÞun ð5:5Þ þ ½ab1 ðxÞ þ b2 ðxÞucn þ an1 þ n2 g Substituting the aggregated HSMC law of the nominal crane system (4. (5.7) work together to achieve the strong robustness of the crane control system against .6). the derivative of V has the form of _ VðtÞ ¼ jS2  gjSj þ S½d0 sgnðSÞ þ an1 þ n2  ð5:8Þ   jS2  gjSj þ S½d0 sgnðSÞ þ jjan1 þ n2 jj _ VðtÞ\0 can be derived from (5. (5.2) and substitute (4.15) into (5.2) into (5.3) into the derivative of V.1) and (4. Then.15) and the compensation law of the unmatched uncertainties in (5.

However. The parameter d0 is defined according to the sum of the two supremums of unmatched uncertainties. and define the control law (5. respectively. At last.2 Stability Analysis Theorem 5. and its function is to compensate the adverse effects of unmatched uncertainties on the system stability. Some physical parameters of the crane are determined in Table 4.2. Then.9).1 and 4. the aggregated sliding surface S at the top layer and the subsystem sliding surfaces s1 and s2 are of asymptotic stability. the final control aw of the uncertain crane system has the form of u ¼ un þ ucn ¼ ueq1 þ ueq2 þ usw þ ucn ð5:9Þ ab1 ðxÞueq1 þ b2 ðxÞueq2  jS  gsgnðSÞ d0 sgnðSÞ ¼  ab1 ðxÞ þ b2 ðxÞ ab1 ðxÞ þ b2 ðxÞ where κ and η are positive and d0  sup jan1 j þ sup jn2 j. h 5. sup jan1 j þ sup jn2 j should be obtained to define the control parameter d0. In the view of mathematics. some numerical results are illustrated. However. and its function is to guarantee the reachability of sliding mode of the top-layer sliding surface. they have private physical significances. the value is hard to know in practice. Proof The proof is very similar to the proofs of Theorems 4. 5.2. In (5.20). They distribute the HSMC law of the nominal crane system and the compensation law of the uncertainties.1 Consider the single-pendulum-type overhead crane system with uncertainties (2.3 Simulation Results To verify the feasibility of the compensator design based on the aggregated structure of sliding surfaces for uncertain crane systems. (4. they can merge each other. two terms include the signum functions.2) and (4. To have the system stability. The initial and desired positions of the trolley are also determined in Table 4. where the parameters of the hierarchical sliding surfaces are kept . The parameter η is inherited from the HSMC law of the nominal crane system.9).2.1.3).1).170 5 Compensator Design Based on Sliding Mode … unmatched uncertainties. The control law based the hierarchical sliding surfaces for the nominal crane system is adopted.1. design the hierarchical sliding surfaces (4.

where the control input u. b Trolley velocity. 4. Some results are displayed in Figs.2.2. there are some slightly residual oscillations because of the existence of uncertainties. the payload angle. and the payload angular velocity are displayed in Fig. Here d0 is selected as 0. In Fig. 5.1c.1 and 5. In the uncertain crane model (2. where rand() is a Matlab command to generate a random number drawn from the standard uniform distribution on the open interval (0. 4. Although the compensator plus controller structure cannot realize the zeroth residual oscillations compared with the nominal crane control system in Sect.2 Compensator Design Based on HSMC 171 unchanged from Sect.1 × [2 × rand()–1]. the subsystem sliding surface s1 at the first layer. d.1a. the control scheme can still resist unmatched uncer- tainties. 4.1. d Payload angular velocity .20).1.14 rad in Fig. 5. where the trolley position. From Fig. only a parameter d0 has to be set according to the two supremums of unmatched uncertainties. obtained by a Simulink model of MATLAB in Appendix A.1 System performance by the compensator design based on the aggregated HSMC method. a Trolley position.5.2 and the coefficients of the switching control law are also kept unchanged from Sect. The maximum payload deviation is about 0. the trolley velocity. 5. both the uncertain terms ξ1 and ξ2 are set by 0. 5. 5.2. c Payload angle. Concerning the compensator design. During the crane transportation.2. the trolley can arrive at the origin from the initial position at about 6 s. the aggregated HSMC law of the nominal crane system works together with the compensation law of uncertainties to realize the robust control of the uncertain crane system.2. In Fig. the Fig.1a–d. The crane performance is illustrated in Fig. The control performance is displayed in Fig. such a small deviation is good enough in practice to increase crane effectiveness and throughout. the compensation law ucn. respectively. 5. 5. 5. 1). 5.1c.

3 Sliding Mode-Based Fuzzy Compensator Design A critical assumption in Sect. In Fig. The fact indicates that the part of the compensation law is very small in the total control law. b Compensation law. However. decrease of the system robustness. a Control input u. The maximum of the absolute compensation law is about 2. and the aggregated sliding surface S at the second layer are shown in Fig.2e. . c Subsystem sliding surface s1 at the first layer.2 is that the system uncertainties must have a known boundary in (5. 5.172 5 Compensator Design Based on Sliding Mode … Fig. and deterioration of the system performance. The lack of knowing the important information may cause several prob- lems such as deficiency of the system stability. To turn SMC-based compensator methods into practical account on the transport control problem of uncertain crane systems. d Subsystem sliding surface s2 at the first layer. the two sliding surfaces at the first layer tend to their subsystems origins along the two subsystem sliding surfaces.2a. 5. 5. In Fig. 5. Thereafter. the sliding mode of the aggregated sliding surface at the top layer is reachable at about 2 s. e Second-layer sliding surface S subsystems sliding surface s2 at the first layer. 5. respectively.2a–e. the boundary value is hard to exactly know in practice.3 N.2 Control performance by the compensator design on basis of the aggregated HSMC method. the maximum of the absolute control input is about 100 N. 5. it is necessary to approximate and compensate the uncertainties.7).

M denotes the total number of rules. The fuzzy logic method has been applied to many practical applications.3. …. 5. b2. mechanical manipulator [4]. where the input vector is a = [a1. fuzzifier. Consider a multi-input multi-output FIS.3. and every element in the vectors a and b is scalar. 5.1 Design of Fuzzy Compensator A fuzzy inference system (FIS) [6] consists of four parts: knowledge base. and the lth rule is described by Fig. The multi-input multi-output if-then rules can be expressed by R ¼ [M l¼1 Rl ð5:10Þ In (5.3 Prototype of the FIS . especially in MEMS triaxial gyroscope system [3]. the output vector b = [b1. and defuzzifier.5. vector voltmeter [5]. The fuzzy logic method is also a possible choice to attack the uncertainties of the SMC-based crane control problem. The structure of this FIS is displayed in Fig.10).3 Sliding Mode-Based Fuzzy Compensator Design 173 The methodology of fuzzy logic has been proven to be effective in dealing with complex nonlinear systems containing uncertainties that are otherwise difficult to model. where the knowledge base is composed of some fuzzy if-then rules and the fuzzy inference engine employs the rules. R denotes a collection of fuzzy rules. am]T 2 <m . 5.3. bn]T 2 <n . …. fuzzy inference engine.1.1 Control Design 5. etc. a2.

11) are the linguistic variables of the fuzzy sets. the product inference engine. h. the approximate values b ¼ ½^n1 . positive small. h.. where p = 1.20) are picked up as the FIS inputs.20) so that the approximate values are def- initely selected as the FIS outputs. ^n2 T . 2. x_ . h. wðaÞ ¼ ½w1 ðaÞ. hÞ ð5:15Þ ^n2 H wðx. ZO. NS. here l = 1. defined by ! ðap  cp Þ2 lAlp ðap Þ ¼ exp  ð5:14Þ 2o2p where cp and op are the center and width of the Gaussian function and p = 1. Associated with the application of the FIS. 2. x_ .12). and bn is Bln : ð5:11Þ Alp and Blq in (5.  bM T l q  2 < is named the parameter vector. PS. From (5.. . am is Alm then b1 is Bl1 . Moreover. denoting negative big. negative small. and positive big. a ¼ ½x. and l PB. hÞ T _ 2 _ is 625 × 1. hÞ T _ ¼ ¼ H wðx. i. n. 4. …. 3. m and q = 1. zero. h The linguistic labels of Ap are chosen as five levels. the purpose is to approximate the uncertainties in the crane dynamics (2. Adopting the singleton fuzzifier mapping. x_ . . ….e. The membership function of Alp is Gaussian.12). NB. . x_ . i. . . the qth output of the FIS has a form of X M bq ¼ bl wl ðaÞ ¼ HT wðaÞ ð5:12Þ q q l¼1 In (5. M.. hÞ . Both the linguistic variables are determined by their membership functions lAlp ðap Þ and lBlq ðbq Þ. and the center-average defuzzifier mapping. . wM ðaÞT 2 <M is the fuzzy basis function vector. where the approximate matrix H is 625 × 2 and wðx. ..e. ^n2 T can be calculated by " # " # ^n1 _ HT1 wðx. The vector Hq is the qth column of an M × n parameter matrix H. 2. b ¼ ½^ n1 . 2. h. …. . . in the crane dynamics (2. x_ . . h.e. where bq is an adaptively M adjustable parameter. where wl ðaÞ is calculated by Qm p¼1 lAl ðap Þ wl ðaÞ ¼ PM Qm p ð5:13Þ l¼1 p¼1 lAlp ðap Þ Hq ¼ ½ b1q . respectively. the system uncertainties have a direct effect on the crane performance so that the state variables _ T.. i.174 5 Compensator Design Based on Sliding Mode … Rl : if a1 is Al1 and a2 is Al2 .

the FIS covers 54 fuzzy rules in its knowledge base. Finally.5. It is proven that fuzzy systems in the form of (5.3 Sliding Mode-Based Fuzzy Compensator Design 175 Provided that an optimal parameter matrix H exists.7) at the top layer of the hierarchical sliding surfaces. One is to cut the number of inputs. adopt the aggregated structure of sliding surfaces. Here. the number of levels is selected as 5 by a trade-off between accuracy and computational burden. Once the number of levels is reduced. The designed FIS is a 4-input. The other is to reduce the number of levels.2-output. the accuracy of the FIS will definitely decrease. the control law can be rewritten by u ¼ un þ ucn ¼ ueq1 þ ueq2 þ usw þ ucn ab1 ðxÞueq1 þ b2 ðxÞueq2  jS  gsgnðSÞ d^0 sgnðSÞ ð5:18Þ ¼  ab1 ðxÞ þ b2 ðxÞ ab1 ðxÞ þ b2 ðxÞ .2 Design of HSMC-Based Controller Consider the uncertain crane system (2. the designed FIS is adopted.12) can approximate continuous function over a compact set to an arbitrary degree of accuracy provided that enough number of rules is given. But the increase will result in computational burden. The uncertain terms have a direct effect on the crane state variables. design the HSMC law for the nominal crane system. the boundary values of the uncertainties are kept unknown so that the control law of the uncertain crane (5. Here the ability of approximation of the FIS is employed to the approximate boundaries of these unknown uncertainties. Considering this assumption. The accuracy of the FIS may increase if the number of levels is increased. where the approximation values of the boundary values are the network outputs. the FIS may not work because it cannot obtain enough information. However. There are two ways to reduce the number of fuzzy rules. and the linguistic variables of the inputs are divided into 5 levels. the matrix has a form of " " #  #  ^n n1    1  H ¼ arg min sup   ð5:16Þ H2v0  ^n n 2  2 where χ0 is a proper set.1) cannot be directly employed here.20). 5. Once the number of inputs is decreased.3. To approximate the boundary values. ρ2]T 2 <21 . and com- pensate the uncertainties by the lumped compensator (5.1. the minimum approximate error vector can be determined by " #   ^n1 n1 q¼  ð5:17Þ ^n2 n2 where ρ = [ρ1. Consequently.

20) into account as a Lyapunov can- didate function 1 1X 2 ~ T Cq H ~q VðtÞ ¼ S2 þ H ð5:20Þ 2 2 q¼1 q ~ q ¼ H  Hq . h. Differentiate S with respect to time t and substitute the uncertain crane dynamics into the derivative of S.176 5 Compensator Design Based on Sliding Mode … In (5. x_ . d^0 determined by the network outputs has the form of d^0 ¼ ja^n1 j þ j^n2 j. S_ ¼ aðc1 x2 þ f1 Þ þ ðc2 x4 þ f2 Þ þ ðab1 þ b2 Þun þ ðab1 þ b2 Þucn þ an1 þ n2 ¼ jS  gsgnðSÞ þ ðab1 þ b2 Þucn þ an1 þ n2 ¼ jS  gsgnðSÞ  d^0 sgnðSÞ þ an1 þ n2 ð5:23Þ . S = αs1 + s2.20) with respect to time t yields X 2 _ VðtÞ ¼ SS_ þ H ~_ q ~ T Cq H ð5:21Þ q q¼1 Further.1).2) and (4. design the hierarchical sliding surfaces (4. and define the control law (5. η and κ are positive scalars.20).2 Consider the single-pendulum-type overhead crane system with uncertainties (2. hÞ _ _ 2 ¼ C1  S  wðx.23) can be deduced from (5. Then.22). hÞ H ð5:19Þ 1 2 where η ≥ |αρ1| + |ρ2| and C1 and C2 are positive.22) can be obtained S_ ¼ a_s1 þ s_ 2 ¼ aðc1 x_ 1 þ x_ 2 Þ þ ðc2 x_ 3 þ x_ 4 Þ ð5:22Þ ¼ aðc1 x2 þ f1 þ b1 u þ n1 Þ þ ðc2 x4 þ f2 þ b2 u þ n2 Þ Consider the control law for the uncertain crane system (5.18). x_ . the aggregated sliding surface S at the top layer is of asymptotic stability if and only if H _ and _ 1 ¼ a1 C1  S  wðx. construct the FIS (5. Then.12).2 Stability Analysis Theorem 5.3). (4.18) and the HSMC law of the nominal crane system. where H q Differentiating (5.3. Equation (5. 5.18). (5. h. Proof Take the positive definite function (5.

25) exist because of η ≥ |αρ1| + |ρ2|. In the sense of Lyapunov. 4.19) are .2.19) of the designed FIS is convergent and the aggregated sliding surface at the top layer is of asymptotic stability. As a result. the subsystem sliding surfaces s1 and s2 are of asymptotic stability. some numerical results are illustrated in Figs. The initial position of the trolley is located at the origin and the desired position is assigned to x = 1 m. h. Proof The proof of Theorem 5. Then. π/4. construct the FIS (5.3 Simulation Results To verify the feasibility and validity of the fuzzy compensator design based on HSMC for uncertain crane systems. −π/4.3. The control law based the hierarchical sliding surfaces for the nominal crane system is adopted. and define the control law (5.20).3 Consider the single-pendulum-type overhead crane system with uncertainties (2. (4. where the parameters of the hierarchical sliding surfaces are kept unchanged from Sect. 0.1. obtained by a Simulink model of MATLAB in Appendix B. h 5. h Theorem 5.1).5.3 is very similar to the proof of Theorem 4.18). The centers of the Gaussian function (5. hÞ q q q¼1 From (5.14) are set by −π/2. Both Γ1 and Γ2 in (5. 5. the crane control system is asymptotically stable. 5. design the hierarchical sliding surfaces (4.6.12).3).13) into (5.2 and the coefficients of the switching control law are also kept unchanged from Sect.2.3 Sliding Mode-Based Fuzzy Compensator Design 177 Consider the FIS output (5. Some physical parameters of the crane are determined in Table 4.21) yields X 2 _ VðtÞ ¼S½jS  gsgnðSÞ  d^0 sgnðSÞ þ an1 þ n2  þ H ~_ q ~ T Cq H q q¼1 X 2  S½jS  gsgnðSÞ  aS½^n1  n1   S½^n2  n2  þ ~_ q ~ T Cq H H ð5:24Þ q q¼1 X 2 ¼ jS2  gjSj þ ½H ~_ q þ SH ~ T Cq H _ ~ T wðx.19). and π/2.12). the adaptive law (5. Substituting (5. x_ . pffiffiffi The width of the Gaussian function is set by 2 π/8. the derivative of V becomes _ VðtÞ ¼ jS2  gjSj þ ðaq1 þ q2 ÞS ð5:25Þ   jS2  gjSj þ ðaq1 þ q2 ÞjSj _ VðtÞ\0 in (5. 4.5 and 5.4.2) and (4.

In the uncertain crane model (2. 5.05 rad.4 System performance by the fuzzy compensator design based on the aggregated HSMC method. where the sliding surface at the top layer S. both the uncertain terms ξ1 and ξ2 are set by 0. From Fig. b Trolley velocity. 5.5. the state trajectories of the two subsystems slide Fig. From Fig.5.4. 5.4c. where the trolley position. 5.4. the subsystem sliding surface s1 at the first layer and the subsystem sliding surface s2 at the first layer are shown in Fig.4.5. respectively.4a–d. the maximum payload deviation is about 0.20). the payload angle has some slight residual oscillations from the angle and angular velocity curves in Fig.1 × [2 × rand() − 1]. the trolley can arrive at the desired position from the origin at about 6 s. Then. the payload angle. Such combinations can realize the strong robustness against uncertainties of crane control systems. 5. 5.178 5 Compensator Design Based on Sliding Mode … set by 0. During the crane movement. where rand() is a MATLAB command to generate a random number drawn from the standard uniform distribution on the open interval (0. the sliding surface at the top layer first arrives at its sliding mode at about 1 s. 5. a Trolley position. d. However. c Payload angle.5a–c. 5. 5. The curves of the aggregated sliding surfaces are displayed in Fig. the aggregated HSMC law of the nominal crane system works together with the fuzzy compensation law of uncertainties against the uncertainties of the crane system. 1). the trolley velocity. Such a small deviation is good enough in real crane applications to increase effectiveness and throughout. The crane performance is illustrated in Fig. d Payload angular velocity . and the payload angular velocity are displayed in Fig. From Fig. respectively.

where the type of radial basis function (RBF) neural networks (NNs) is often employed as a universal approximator [8]. only the reachability of the sliding surface at the top layer is guaranteed. 5. The neural network types vary a lot. 5. b Subsystem sliding surface s1 at the first layer. and the compensator output ^n2 are displayed in Fig.6b.2 and 5. 5. u has to switch in high frequency to resist the system uncertainties. From Fig.3. and they use weights to change the parameters of the throughput and the varying connections to the neu- rons. In [9]. However.6 shows the control performance by the fuzzy compensation law based on the aggregated HSMC method.4 Sliding Mode-Based Neural Compensator Design Neural networks [7] use algorithms in their programming.5 Hierarchical sliding surfaces. Figure 5.14) till the compensator outputs tend to be stable. a Sliding surface S at the top layer. the maximum control input is about 40 N.6a–c. a gradient-type method on basis of RBF NNs is proposed to deal with the . From Fig. c Subsystem sliding surface s2 at the first layer along their private sliding surface and converge to their subsystem origins.6a. 5. the two compensators can adaptively track the system uncertainties according to the adaptive law (5. the compensator output ^ n1 . The sliding modes of the subsystem sliding surfaces cannot be reached at finite time and their sliding modes are asymptotic.3 Sliding Mode-Based Fuzzy Compensator Design 179 Fig. respectively. As proven in Theorems 5. c. 5.5. where the control input u.

6 Control performance by the fuzzy compensator design based on the aggregated HSMC method.1 Design of Neural Compensator RBF networks are a kind of three-layer feed-forward networks. where the mapping from the input layer to the output layer is inherently nonlinear but the mapping from the hidden layer to the output layer is linear. Weight update formulas of the networks are deduced from Lyapunov direct method. 5.4. Since RBF NNs have the ability to approximate complex nonlinear mapping directly from input–output data with a simple topological structure. . Provided that the crane uncertainties have an unknown boundary. 5. c Compensator output ^ n2 dead band nonlinearity. so the weight convergence and system stability are simultaneously guaranteed in the sense of Lyapunov. such kind of NNs is adopted here to realize the compensator design [10].4 are designed to compensate the crane uncertainties. b Compensator output ^ n1 . the RBF NNs in Sect.4.1.1 Control Design 5. 5. a Control input u.180 5 Compensator Design Based on Sliding Mode … Fig.

h. 5. Associated with the application of the RBF networks. Both cp and οp are predefined.5.20) so that the approximate values are definitely selected as the network output. cp 2 <m is the center vector of the pth Gaussian function.20). where q = 1. the network output of the ith network can be calculated by ^ni ðx.7.20).7 Prototype of the RBF networks The prototype of the designed RBF networks is illustrated in Fig. In Fig. There are l neurons in the hidden layer. x1 Þ xT1 hðxÞ n¼ ¼ ð5:28Þ ^n2 ðx. xi Þ ¼ xT hðxÞ ð5:26Þ i where xi 2 <l is the weight vector of the ith RBF network. Consider the uncertain crane system (2. x_ . 2. ½x. hðxÞ 2 <l is the Gaussian function vector and the pth element hp(x) of the vector hp(x) is defined by ! jjx  cp jj2 hp ðxÞ ¼ exp  ð5:27Þ 2o2p In (5. At the output layer. h From Fig. l. …. two RBF networks have to be employed and the ith network output is described by yi ¼ ^ni (i = 1. there are m neurons at the input layer and each element xq is employed as an input element. …. only 1 neuron is assigned and the network output is written by y. Moreover. x2 Þ xT2 hðxÞ . m. In other words. the purpose is to approximate the uncertainties in the crane dynamics (2. To approximate the two uncertain terms in (2. the system uncertainties have a direct effect on the crane performance so that the state variables x ¼ _ T in the crane dynamics (2. 5.20) are picked up as the network inputs.7. the designed RBF networks have m inputs and 1 output. 5. the vector n describing the approximate values can be written by " # " # ^n1 ðx.27). p = 1. 5. 2). οp is scalar.4 Sliding Mode-Based Neural Compensator Design 181 Fig. indicating the width of the pth Gaussian function.7. Since the compensator has two RBF networks.

29). there exists an optimal weight vector xi such that the network output satisfies  T  x hðxÞ  sup jni j ¼ eðxÞ\ei ð5:30Þ i 0 Assumption 5. To approximate the boundary values.4.4 Consider the single-pendulum-type overhead crane system with uncertainties (2.20).26).20).182 5 Compensator Design Based on Sliding Mode … 5.7) at the top layer of the hierarchical sliding surfaces. design the hierarchical sliding surfaces (4.2 Design of SMC-Based Controller Consider the uncertain crane system (2.3).2) and (4.32).2 The boundary value sup jni j minus the norm of the system uncertainties satisfies the following formula sup jni j  jni j [ ei1 [ ei0 ð5:31Þ Theorem 5.1). and com- pensate the uncertainties by the lumped compensator (5. .1. 5.1) cannot be directly employed here. design the HSMC law for the nominal crane system. However.1 For either of the designed RBF networks. adopt the aggregated structure of sliding surfaces. and define the control law (5.2 Stability Analysis Assumption 5.4. then the aggregated sliding surface S at the top layer is of asymptotic stability. construct the RBF neural networks (5. x_ 1 ¼ aC1 1  jSj  hðxÞ ð5:32Þ x_ 2 ¼ C1 2  jSj  hðxÞ where C1 and C2 are positive constants. (4. both of the designed RBF networks are adopted and the control law can be rewritten by u ¼ un þ ucn ¼ ueq1 þ ueq2 þ usw þ ucn ð5:29Þ ab1 ðxÞueq1 þ b2 ðxÞueq2  jS  gsgnðSÞ d^0 sgnðSÞ ¼  ab1 ðxÞ þ b2 ðxÞ ab1 ðxÞ þ b2 ðxÞ where d^0 ¼ ja^n1 j þ j^n2 j. the boundary values of the uncertainties are kept unknown so that the control law of the uncertain crane (5. If the update formulas are defined by (5.

38) becomes h i _  S jS  gsgnðSÞ  að^n1  n1 Þ  ð^n2  n2 Þ VðtÞ  aðx1  x1 ÞT hðxÞjsj  ðx2  x2 ÞT hðxÞjsj ð5:39Þ ¼ jS2  gjSj  aðe11  e10 ÞjSj  ðe21  e20 ÞjSj . Substituting (5.36) can be obtained S_ ¼ a_s1 þ s_ 2 ¼ aðc1 x_ 1 þ x_ 2 Þ þ ðc2 x_ 3 þ x_ 4 Þ ð5:36Þ ¼ aðc1 x2 þ f1 þ b1 u þ n1 Þ þ ðc2 x4 þ f2 þ b2 u þ n2 Þ Consider the control law for the uncertain crane system (5.5.4 Sliding Mode-Based Neural Compensator Design 183 Proof Take the positive definite function (5.36). (5. Equation (5. Differentiate S with respect to time t and substitute the uncertain crane dynamics into the derivative of S. x ~ i ¼ xi  xi x ð5:34Þ Differentiating (5. (5.32).33) with respect to time t yields X 2 _ VðtÞ ¼ SS_  ~ Ti x_ i Ci x ð5:35Þ i¼1 Further. Then.33).37) into (5.33) into account as a Lyapunov can- didate function 1 X2 C1 T VðtÞ ¼ S2 þ ~ x x ~i ð5:33Þ 2 i¼1 2 i ~ i is determined by In (5.28) of the RBF neural networks. S = αs1 + s2.29) and the HSMC law of the nominal crane system.35) yields X 2 _  S½jS  gsgnðSÞ  að^n1  n1 Þ  ð^n2  n2 Þ  VðtÞ ~ Ti x_ i Ci x ð5:38Þ i¼1 From (5.37) can be deduced from (5. S_ ¼ aðc1 x2 þ f1 Þ þ ðc2 x4 þ f2 Þ þ ðab1 þ b2 Þun þ ðab1 þ b2 Þucn þ an1 þ n2 ð5:37Þ ¼ jS  gsgnðSÞ þ ðab1 þ b2 Þucn þ an1 þ n2 ¼ jS  gsgnðSÞ  d^0 sgnðSÞ þ an1 þ n2 Consider the outputs (5.

20). From Fig. the subsystem sliding surfaces s1 and s2 are of asymptotic stability.8b. Such combina- tions can realize the strong robustness against uncertainties of crane control systems. The initial weights of the two neural networks are designed as random numbers on the open interval (0. 1). construct the RBF networks (5. h 5.10. design the hierarchical sliding surfaces (4. 5.8a.2 rad. (4.3). 5.3 Simulation Results To verify the feasibility and validity of the neural compensator design based on HSMC for uncertain crane systems. respectively. the payload angle has some residual oscillations because of the adverse effects of uncertainties. the payload angle. and define the control law (5.39) exists.8a–d. 1). the subsystem sliding surface s1 at the first layer and the subsystem sliding surface s2 at the first layer are shown in Fig. and the payload angular velocity are displayed in Fig. 4. where the trolley position. Consequently.8.2. From Fig.27) [11] are set by 1.2 and the coefficients of the switching control law are also kept unchanged from Sect. The initial position of the trolley is located at x = 2 m and the desired position is assigned to the origin. In the sense of Lyapunov.5 Consider the single-pendulum-type overhead crane system with uncertainties (2. The crane performance is illustrated in Fig. The maximum angular deviation is about 0. the aggregated HSMC law of the nominal crane system works together with the neural network-based compensation law of uncertainties against the uncertainties of the crane system. h Theorem 5. In the uncertain crane model (2.2. VðtÞ\0 in (5.4. The curves of the aggregated sliding surfaces are displayed in Fig. the crane control system is asymptotically stable.8. where the sliding surface at the top layer S.32) of the designed RBF networks is convergent and the aggregated sliding surface at the top layer is of asymptotic stability. . the adaptive law (5. Then.20).9a–c. d. 5. the trolley velocity.1).8. Proof The proof of Theorem 5. where the parameters of the hierarchical sliding surfaces are kept unchanged from Sect.5 is very similar to the proof of Theorem 4. From Fig.1 × [2 × rand() − 1]. 5. 4.2) and (4. obtained by a Simulink model of MATLAB in Appendix C. 5. Some physical parameters of the crane are determined in Table 4.1 and 5. 5.19) are set by 1079.1. 5. The control law based the hierarchical sliding surfaces for the nominal crane system is adopted. the trolley can arrive at the origin from the initial position at about 6 s. where rand() is a MATLAB command to generate a random number drawn from the standard uniform distribution on the open interval (0. Both Γ1 and Γ2 in (5. 5.29).26).9 and 5. both the uncertain terms ξ1 and ξ2 are set by 0.2. Both the centers and width of the Gaussian function (5.9. some numerical results are illustrated in Figs. 5.184 5 Compensator Design Based on Sliding Mode … _ Concerning Assumptions 5.

9 Hierarchical sliding surfaces. a Sliding surface S at the top layer. c Subsystem sliding surface s2 at the first layer .8 System performance by the neural compensator design based on the aggregated HSMC method. b Trolley velocity. 5. d Payload angular velocity Fig.5.4 Sliding Mode-Based Neural Compensator Design 185 Fig. c Payload angle. a Trolley position. 5. b Subsystem sliding surface s1 at the first layer.

As proven in Theorems 5.32) till the compensator outputs tend to be stable. c.4 and 5. where the control input u. 5. u has to switch in high frequency to resist the system uncertainties. the state trajectories of the two subsystems slide along their private sliding surface and converge to their subsystem origins. 5. Compared with the results in Fig. The sliding modes of the subsystem sliding surfaces cannot be reached at finite time and their sliding modes are asymptotic. 5.5. 5. only the reachability of the sliding surface at the top layer is guaranteed. respec- tively. the results in Fig. Then.10a–c. the two neural-network-based compensators can adaptively track the system uncertainties according to the adaptive law (5. and the compensator output ^n2 are displayed in Fig. the compensator output ^ n1 .10 shows the control performance by the neural compensation law based on the aggregated HSMC method.186 5 Compensator Design Based on Sliding Mode … Fig.10 Control performance by the neural compensator design based on the aggregated HSMC method. . 5. c Compensator output ^n2 respectively.10a. c. However. From Fig.9. a Control input u. the sliding surface at the top layer first arrives at its sliding mode at about 1. However. the neural-network-based compensators have the zeroth outputs as t tends to infinite. From Fig. The fuzzy logic-based compensator design has constant outputs as approximation values of uncertainties. the maximum control input is about 50 N. b Compensator output ^ n1 . From Fig. c illustrate that the neural-network-based compensators have different outputs. Figure 5.10b.6b. 5.5 s. 5.10b.

case 1.DirFeedthrough = 0.str. x0 = [2 0 0 0].str. case {2.4. sizes.x. sizes.x.num2str(flag)]).x.ts]=mdlInitializeSizes.NumSampleTimes = 0.x0.ts]=mdlInitializeSizes sizes = simsizes. case 0. sys=mdlOutputs(t.Appendices 187 Appendices A Simulink Model to Plot Figs. case 3.NumInputs = 1.str. 5.str = [].u) .x.NumContStates = 4.x0. sizes.u.NumDiscStates = 0.NumOutputs = 4.ts = []. sizes. % at least one sample time is needed sys = simsizes(sizes).u).ts] = sfuntmpl(t. sys=mdlDerivatives(t. sizes.x0.9} sys=[]. function sys=mdlDerivatives(t. end function [sys. otherwise error(['Unhandled flag = '.flag) switch flag. function [sys. m.2 Plant program: SPCrane. [sys.1 and 5. sizes.u).

2. b1 =D*inv(A*D-B*C).NumContStates = 0. sys(3)=x(4). x0 = [].NumOutputs = 5.x0.flag) switch flag.81.u) sys = x.ts] = sfuntmpl(t. end function [sys.l=1.D=l.B=(m*l*cos(u(3))).ts]=mdlInitializeSizes sizes = simsizes. A=(M+m). case 0.05.str.1*(2*rand()-1).m=5. sizes.x.str. A=(M+m).str.str = []. sizes. sys=mdlOutputs(t. sys(1)=x(2).u) % parameter of system g=9.NumDiscStates = 0.C=(cos(u(3))).NumSampleTimes = 0. case {1.ts = [].4. Controller program: CtrPCmp.x. f2 =(m*l*x(4)^2*sin(x(3))*C+g*A*sin(x(3)))*inv(B*C-A*D). sizes.M=37.188 5 Compensator Design Based on Sliding Mode … %parameter g=9.u. m function [sys.ts]=mdlInitializeSizes. sizes.m=5.05. % parameter of controller . sys(2)=((u(1)+m*l*x(4)^2*sin(x(3)))*D+g*B*sin(x(3)))*inv(A*D- B*C)+0.D=l. b2 =C*inv(B*C-A*D). sizes.x0.x.B=(m*l*cos(x(3))).x.C=(cos(x(3))).E=(m*l*sin(u(3))).32.NumInputs = 4.num2str(flag)]). function sys=mdlOutputs(t.1*(2*rand()-1). % at least one sample time is needed sys = simsizes(sizes).u). otherwise error(['Unhandled flag = '. [sys. f1 =(m*l*x(4)^2*sin(x(3))*D+g*B*sin(x(3)))*inv(A*D-B*C).l=1.DirFeedthrough = 1. function sys=mdlOutputs(t. sizes. case 3.x0.M=37.32.9} sys=[]. sys(4)=((u(1)+m*l*x(4)^2*sin(x(3)))*C+g*A*sin(x(3)))*inv(B*C- A*D)+0.81.

f2 =(m*l*u(4)^2*sin(u(3))*C+g*A*sin(u(3)))*inv(B*C-A*D). dd0=0.3.e(3)=u(3). sys(3)=s2. uslaw=-num*inv(den)+ds*inv(den)-dd0*inv(den). e(1)=(u(1)). k=3. a =-2. s=S2. num=a*(afa(1)*u(2)+bata(1)*f1)+(afa(2)*u(4)+bata(2)*f2).e(4)=u(4). S1=s1. S2=a*S1+s2. 5.ita=0.afa =[0.7 8.5 and 5. B Simulink Model to Plot Figs.4.2]. b2 =C*inv(B*C-A*D). sys(1)=uslaw. den=a*bata(1)*b1+bata(2)*b2.e(2)=u(2). sys(5)=-dd0*inv(den). sys(2)=s1. ds=-k*s-ita*sign(s).2. sys(4)=S2.6 . f1 =(m*l*u(4)^2*sin(u(3))*D+g*B*sin(u(3)))*inv(A*D-B*C). s2=afa(2)*e(3)+bata(2)*e(4).bata=[1 1].Appendices 189 %e=x-xd. 5.1. b1 =D*inv(A*D-B*C). s1=afa(1)*e(1)+bata(1)*e(2).

Controller program: Ctr_FuzzyCmp.1)]. sys=mdlDerivatives(t.4.1*ones(1250. sizes. case 3.ts]=mdlInitializeSizes sizes = simsizes.9} sys=[].ts]=mdlInitializeSizes.ts] = sfuntmpl(t. Chap.u).m file in Appendix A. u2(l2)=exp(gs2).str. sys=mdlOutputs(t. for l2=1:1:5 gs2=-[(q2+pi/2-(l2-1)*pi/4)/(pi/4)]^2. function sys=mdlDerivatives(t. 5.u) q1=u(1). sizes.x. case {1. sizes.NumContStates = 1250.x. sizes.x. U1(l11)=exp(gs11). sizes. str = [].DirFeedthrough = 1.dq2=u(4). ts = [0 0]. x0 = [0.dq1=u(2).NumDiscStates = 0. fsd1=0. case 1.flag) switch flag.num2str(flag)]).x0. The file is very similar to the SPCrane. end . end function [sys.u. u1(l1)=exp(gs1). m function [sys.190 5 Compensator Design Based on Sliding Mode … Plant program: SPCrane.NumOutputs = 6.q2=u(3). % at least one sample time is needed sys = simsizes(sizes). m. end fsd2=0. otherwise error(['Unhandled flag = '.str.str.NumInputs = 6.u).x. end for l11=1:1:5 gs11=-[(dq1+pi/2-(l11-1)*pi/4)/(pi/4)]^2. [sys.2. case 0. sizes.x0.NumSampleTimes = 1. for l1=1:1:5 gs1=-[(q1+pi/2-(l1-1)*pi/4)/(pi/4)]^2.x0.

s=alpha1*s1+alpha2*s2.l=1.alpha1=-3.81.x. end fs2=GG'/(fsd2+0.BB=AA(:). end end CC=fsu2'.DD=CC(:). s2=e4+lamad2*e3. %state error.001). S1=1/Gama1*s*fs1.GG=FF(:). S2=1/Gama2*s*fs2.lamad2=10.alpha2=1.lamad2=8.32. lamad1=1.05. U2(l22)=exp(gs22).7.5. end function sys=mdlOutputs(t.alpha1=-2.001).1.e4=u(4). end fs1=GG'/(fsd1+0. for i=1:1:625 sys(i)=S1(i). end for j=626:1:1250 sys(j)=S2(j-625). % parameter of system g=9. for l2=1:1:5 for l22=1:1:5 fsu2=u2'*U2. end for l1=1:1:5 for l11=1:1:5 fsu1=u1'*U1. for L2=1:1:625 fsd2=fsd2+EE(L2).2.3. EE=BB*DD'.u) ita=0. s1=e2+lamad1*e1.Gama2=0. end end AA=fsu1'. for L1=1:1:625 fsd1=fsd1+EE(L1).lamad1=0.alpha2=1.m=5. Gama1=0.k=1. .FF=EE'.e2=u(2).Appendices 191 for l22=1:1:5 gs22=-[(dq2+pi/2-(l22-1)*pi/4)/(pi/4)]^2.M=37. e1=(u(1))-1.5. e3=u(3).

e3=u(3). f2=(D*C*(u(4))^2+A*g*sin(u(3)))/(B*C-A*l).e2=u(2). end fsd2=0. b1=1/(A-B*C/l). end for l22=1:1:5 gs22=-[(dq2+pi/2-(l22-1)*pi/4)/(pi/4)]^2. end end AA=fsu1'. for l1=1:1:5 gs1=-[(q1+pi/2-(l1-1)*pi/4)/(pi/4)]^2.DD=CC(:). U1(l11)=exp(gs11).dq2=u(4). s1=e2+lamad1*e1.e4=u(4). fsd1=0. for l2=1:1:5 for l22=1:1:5 fsu2=u2'*U2. f1=(D*(u(4))^2+C*D*g/l)/(A-B*C/l). end for i=1:1:625 thta2(i. s2=e4+lamad2*e3. end for l1=1:1:5 for l11=1:1:5 fsu1=u1'*U1.1)=x(i). e1=(u(1))-1. ds=ita*sign(s)+k*s.1)=x(i+625).dq1=u(2). end end CC=fsu2'.q2=u(3). . U2(l22)=exp(gs22).B=(m*l*cos(u(3))).BB=AA(:). for l2=1:1:5 gs2=-[(q2+pi/2-(l2-1)*pi/4)/(pi/4)]^2.192 5 Compensator Design Based on Sliding Mode … A=(M+m). C=(cos(u(3))). end q1=u(1). u2(l2)=exp(gs2). %state error. for i=1:1:625 thta1(i. end for l11=1:1:5 gs11=-[(dq1+pi/2-(l11-1)*pi/4)/(pi/4)]^2. u1(l1)=exp(gs1).D=(m*l*sin(u(3))). s=alpha1*s1+alpha2*s2. b2=C/(B*C-A*l).

%adopt them to make comparison with the desired fuzzy compen- sator con_law=inv(alpha1*b1+alpha2*b2)*(-alpha1*f1-alpha2*f2- abs(alpha1*Fp1)*sign(s)-abs(alpha2*Fp2)*sign(s).9 and 5. %two estimated boundaries of uncertainties %Fp2=0. for L2=1:1:625 fsd2=fsd2+EE(L2). Fp2=thta2'*alpha1*fs2'. sys(3)=Fp1.GG=FF(:).. sys(6)=s2. end fs2=GG'/(fsd2+0.10 . -alpha1*lamad1*e2-alpha2*lamad2*e4-ds). for L1=1:1:625 fsd1=fsd1+EE(L1).001). %out of compensator sys(5)=s1. %Fp1=0.1.Appendices 193 EE=BB*DD'.. 5.1.8. 5. C Simulink Model to Plot Figs. sys(1)=con_law. Fp1=thta1'*alpha1*fs1'.001). FF=EE'. sys(2)=s. %out of compensator sys(4)=Fp2. end fs1=GG'/(fsd1+0.

lamad1=0.ts]=mdlInitializeSizes.num2str(flag)]). end . Chap.u). sizes.u) we1=1*ones(4.1). case 3.ts = [].DirFeedthrough = 1.str. m.j))^2/(2*bb2(j)^2)). function sys=mdlDerivatives(t.str. ds=ita*sign(s)+k*s. sys=mdlDerivatives(t.ts]=mdlInitializeSizes sizes = simsizes.6). sizes.u). case 0.2.x0.NumSampleTimes = 0. sizes. 5.7.6 2.flag) switch flag. otherwise error(['Unhandled flag = '. end function [sys. x0 = [1*rand(12. case 1.we2=1*ones(4. case {2. The file is very similar to the SPCrane.alpha1=-2. [sys.lamad2=8.j))^2/(2*bb1(j)^2)).2.1)].NumContStates = 12. s1=u(2)+lamad1*u(1).NumInputs = 4. xstate=[u(1). % at least one sample time is needed sys = simsizes(sizes). m function [sys.str = []. sys=mdlOutputs(t.u(3).k=3.bb2=1*ones(6.6).alpha2=1.x.NumOutputs = 6.3.1).x. end for j=1:6 h2(j)=exp(-norm(xstate-we2(:.m file in Appendix A. sizes.str.u. s=alpha1*s1+alpha2*s2.x0.x.ts] = sfuntmpl(t. Controller program: Ctr_NeuralCmp. sizes.x0.NumDiscStates = 0. s2=u(4)+lamad2*u(3).194 5 Compensator Design Based on Sliding Mode … Plant program: SPCrane. sizes. for j=1:6 h1(j)=exp(-norm(xstate-we1(:. ita=0.u(2).x.3 -8.9} sys=[].7 1].4. % s=c*xstate. % c=[1.bb1=1*ones(6.u(4)].

xite2=10^81. num=alpha1*lamad1*u(2)+lamad2*u(4)+alpha1*f1+f2.7 1].6 2.6). s=alpha1*s1+alpha2*s2. w2=[x(7).05.1).lamad1=0.2.x(8). xstate=[u(1).E=(m*l*sin(u(3))).D=l. % eta=0.3 -8.u(4)]. eq1=0. w1=[x(1). s2=u(4)+lamad2*u(3).B=(m*l*cos(u(3))). A=(M+m). f1 =(m*l*u(4)^2*sin(u(3))*D+g*B*sin(u(3)))*inv(A*D-B*C).u(3).m=5.u) we1=10*ones(4.u(2).2.Appendices 195 eq0=0.bb2=1*ones(6.M=37.3. % parameter of controller %e=x-xd.20*ones(6. den=alpha1*b1+b2.alpha2=1.x.alpha1=-2.001. b2 =C*inv(B*C-A*D).1). for i=1:6 sys(i)=xite1*abs(s)*h1(i).bb1=1*ones(6.l=1.7.lamad2=8. % s=c*u. % parameter of system g=9. b1 =D*inv(A*D-B*C). % b2=0.x(5).1). f2 =(m*l*u(4)^2*sin(u(3))*C+g*A*sin(u(3)))*inv(B*C-A*D).x(10).x(2).e(4)=u(4). % we1=1*ones(4.x(3).e(3)=u(3). end for i=1:6 sys(i+6)=xite2*abs(s)*h2(i).C=(cos(u(3))).x(9).x(6)]. ds=ita*sign(s)+k*s.01.32. s1=u(2)+lamad1*u(1).x(11).e(2)=u(2).1).6). xite1=alpha1*10^81. % we2=1*ones(4. % k=10.20*ones(6.x(4).002.6).k=3. for j=1:6 .we2=10*ones(4.x(12)]. ita=0. % b1=0.6). end function sys=mdlOutputs(t. e(1)=(u(1)).81. % c=[1.

Qian DW. Up2=w2'*h2'. Rakshit A (2011) A reinforcement-learning-based fuzzy compensator for a microcontroller-based frequency synthesizer/vector voltmeter. Liu XJ (2011) A robust sliding mode controller based on RBF neural networks for overhead crane systems with uncertain dynamics. Neural Comput Appl 22(2):279–286 11. Academic Press. Yi JQ. ds=-k*s-(ita)*sign(s). Sun FC. Fei JT. IEEE Trans Syst Man Cybern. Liu XJ. Baughman DR. Neural Comput 3(2):246–257 9. sys(5)=s1. Chatterjee A. Zhao DB. Proc Inst Mech Eng Part I J Syst Control Eng 223 (6):785–795 2. Qian DW. sys(2)=Up1. Sandberg IW (1991) Universal approximation using radial-basis-function networks. Sarkar G. Thro E (2014) Fuzzy logic: a practical approach. Liu YA (2014) Neural networks in bioprocessing and chemical engineering. Qian DW. Int J Nonlinear Sci Numer Simul 15(3– 4):181–188 5. USA 8.j))^2/(2*bb1(j)^2)). ICIC Express Lett 5(6):1995– 2000 . Song Q (2010) A robust training algorithm of discrete-time MIMO RNN and application in fault tolerant control of robotic system. end Up1=w1'*h1'. Wu YL. USA 7. Neural Comput Appl 19(7):1013–1027 10. Academic Press. Yi JQ (2009) Robust sliding mode control for a class of underactuated systems with mismatched uncertainties. Lin CM. IEEE Trans Control Syst Technol 13(4):593–598 3. B Cybern 42(6):1599–1607 4.196 5 Compensator Design Based on Sliding Mode … h1(j)=exp(-norm(xstate-we1(:. Yi JQ. References 1. sys(6)=s2. end for j=1:6 h2(j)=exp(-norm(xstate-we2(:. Mon YJ (2005) Decoupling control by hierarchical fuzzy sliding-mode controller. Singh HP (2014) Simulation of neural network based adaptive compensator control scheme for multiple mobile manipulators with uncertainties. uslaw=-num*inv(den)+ds*inv(den)-(alpha1*Up1+Up2)*sign(s)*inv(den). sys(3)=Up2. Zheng JC. sys(4)=s. IEEE Trans Instrum Meas 60(9):3120–3127 6. Zhou J (2012) Robust adaptive control of MEMS triaxial gyroscope using fuzzy compensator. Park J. sys(1)=uslaw.j))^2/(2*bb2(j)^2)). Liu XJ (2013) Neural sliding-mode load frequency controller design of power systems. McNeill FM.

By capturing the characteristic. Concerning the SMC-based design for overhead crane systems. entitled aggregated HSMC. terminal SMC. Such an inherent property challenges their control design. The physical characteristic is that an overhead crane is composed of several subsystems.   Keywords Hierarchical sliding mode control Overhead crane Transport control 6. Compared with the traditional single-layer sliding surface of several SMC methods. DOI 10. Numerical simulations show the © Springer-Verlag Berlin Heidelberg 2016 197 D.1007/978-3-662-48417-3_6 . Yi. Hierarchical Sliding Mode Control for Under-actuated Cranes. 1st-order SMC.. the methodology of hierarchical sliding mode control (HSMC) is presented.e. This methodology covers three basic types of hierarchical structures. This book investigates the three kinds of HSMC methods for single-pendulum-type overhead crane systems. this book focuses on the structure design of sliding surfaces. Sliding mode control (SMC) is recognized as one of the efficient tools to design robust controllers for complex nonlinear systems. it is one of the most active fields of research in control community. Qian and J.Chapter 6 Conclusions and Open Problems Abstract This chapter presents preliminary results on the development and application of hierarchical sliding mode control (SMC) methods to the transport control design of overhead crane systems and draws some concluding remarks. etc.1 Conclusions Overhead crane systems are under-actuated because they have a lower number of control inputs than the number of degrees of freedom to be controlled. incremental HSMC and combing HSMC. integral SMC. i. For such a wide research area. the HSMC methodology has significantly contributed to reveal the sliding motion by the phase-plane method because the hierarchical sliding surfaces consists of a series of 2nd-order sliding surfaces. 2nd-order SMC. Such hierarchical surfaces give an insight into the sliding motion. and summarizes open problems for future research.. The HSMC methodology employs the hier- archical structure design of sliding surfaces for overhead cranes.

To guarantee the crane control systems is of asymptotic stability in the presence of uncertain. Although SMC is of invariance again matched uncertainties. This book touches this topic and adopts the structure of controller plus com- pensator to deal with the transport control problem of uncertain overhead cranes. Any real overhead crane systems are subject to uncertainties.198 6 Conclusions and Open Problems crane state variables. the compensator design on basis of hierarchical sliding surfaces is touched in this book. The applications of HSMC can also be extended to double-pendulum-type overhead cranes. the assumption is too strict to satisfy under some conditions. This book assumes that the payload angle is measurable such that payload angle and its angular velocity are directly adopted for control design without more explanations. uncertainties involved in crane dynamics cover both matched and unmatched uncertainties. These uncertainties challenge the design of HSMC meth- ods. several important issues and possible future research directions are drawn: Hierarchy. The theoretical analysis and numerical simulations demonstrate the feasibility and robustness of the designed methods against uncertainties. Payload angle. these uncertainties are not accurately modeled. the lumped compensator structure is adopted. This book analyzes that overhead crane systems are subject to uncertainties. However. . Uncertainties can be categorized by matched uncertainties and unmatched ones. This book attempts the technologies of fuzzy logic and neural networks to make the assumption mild. It is meaningful and significant to investigate transport control problems of over- head cranes under the limited measurable state variables. The main goal of implementing the HSMC methods again uncertainties is to guarantee the crane control systems is of asymptotic stability in the presence of uncertain overhead crane systems. a common assumption is that uncertainties should have a known boundary. The hierarchical structure of sliding surfaces offers two possible compensator structures. But payload angle may be hard to obtain in practice [3]. To suppress the uncertainties and compensate them. However. the hierarchical sliding surfaces. To simplify the problem. Current contributions about HSMC are hierarchical structures of sliding surfaces. control systems are also of hierarchy. 6. it is necessary to investigate this field. This indicates there are still some spaces for current HSMC methods to improve their intelligence.2 Extensions and Open Problems Control of overhead crane systems by HSMC is open and it is difficult to include all the works in this book. Modeling of uncertainties. To overcome the adverse effects of unmatched uncertainties. the control performance and the sliding modes based on HSMC. Such hierarchy is only located at the controller level rather than at the control-system level. Finally. Pointed out by Lima and Saridis [2]. Some numerical results are also illustrated. we believe that this book can provide readers with the most up-to-date knowledge in the field. However.

2 Extensions and Open Problems 199 Parameter tuning. USA 2. it is possible to improve HSMC methods by employing evolutional algorithms on this aspect. Taylor & Francis. The feasibility and validity of HSMC methods have been demonstrated by overhead cranes. Since there are several controller parameters for each control method. Utkin V. To dem- onstrate HSMC methods. Lima PU. this book only takes overhead cranes as an example to illustrate the control design and stability analysis. There are three types of cranes. It is possible and expected to extend HSMC methods to the other two types of cranes. Since the other two types of cranes are described by non-Cartesian coordinates. Guldner J. Singhose W (2009) Command shaping for flexible systems: a review of the first 50 years. But parameter tuning is a time-consuming business. In this book. World Scientific. Shi J (2009) Sliding mode control in electromechanical systems. some technical details about HSMC design remain problematic and challenging. 2nd edn. controller parameters are tuned after trial and error. USA 3. Extensions of other types of cranes. Saridis GN (1996) Design of intelligent control systems based on hierarchical stochastic automata. Controller parameters have direct effects on the system sta- bility and robustness. Int J Precis Eng Manuf 10(4):153–168 .6. References 1.