35 views

Uploaded by Alok Pandey

optimization pf antenna parameters using soft computation

- 4_1996-Fuzzy Logic=Computing With Words
- redes neuronales
- genetic fuzzy logic
- BrainSlug- Transparent Acceleration of Deep Learning Through Depth-First Parallelism
- Intelligent Decision Support Systems.ppt
- bell2
- A PREDICTIVE MODEL FOR MAPPING CRIME USING BIG DATA ANALYTICS.pdf
- Optimisation of Composite Boat Hull Structures
- 2-383-15051090329-11 (1)
- a Neuro-Fuzzy Controller for Collaborative Applications in Robotics Using LabVIEW
- Week11 Fuzzy Logic Tutorial
- prl16rrr
- A Multiple Level MIMO FL Based Intelligence For Multi Agent Robot System
- Paper 14
- Medical Diagnosis Classification
- 10.1.1.21.7030
- Neutrosophic Logic Approaches Applied to ”RABOT” Real Time Control
- Scheduling Jobs on Parallel Machines Applying Neural
- Determination of Combustion Parameters Using Engine Crankshaft Speed
- Research on Active Control of Diesel Engine Based on Wavelet Neural Network

You are on page 1of 4

Associate Professor, Dept. of ECE, SIIT,Jaipur 302029 alok.pandey@ieee.org

1

Abstract Co active neural networks & Fuzzy Inference System (CANFIS) are electronic systems which can be trained to remember behavior of a modeled structure in given operational points, and which can be used to approximate behavior of the structure out of the training points. These approximation abilities of neural nets are demonstrated on modeling a microstrip antenna. Parameters are turned to the accuracy and to the efficiency of neural models. The association of neural models and fuzzy algorithms in microstrip antenna design is discussed in this paper. The CANFIS (Co-Active Neuro-Fuzzy Inference System) model integrates adaptable fuzzy inputs with a modular neural network to rapidly and accurately approximate complex functions. Fuzzy inference systems are also valuable as they combine the explanatory nature of rules (membership functions) with the power of "black box" neural networks. Keywords: Co active neural networks & Fuzzy Inference System, microstrip antennas, modeling, optimization

design and analysis of microstrip antennas in rectangular form. In this work, the width and the length for designing of rectangular microstrip antennas are obtained by using new method based on artificial neural networks. The artificial neural networks are developed from neurophysiology by morphologically and computationally mimicking human brains. Although the precise operation details of artificial neural networks are quite different from human brains, they are similar in three aspects: they consist of a very large number of processing elements each neuron connects to a large number of other neurons, and the functionality of networks is determined by modifying the strengths of connections during a learning phase. The purpose of this article is to provide an overview of recent developments in the design and analysis of microstrip antenna using neural network and this is illustrated by taking an illustrative example of microstrip antenna design using ANN. 2. CANFIS Architecture

1.

Introduction

Microstrip antennas due to their many attractive features have drawn popularity among researchers and for industrial application. In high-performance spacecraft, aircraft, airship, aerostat, missile, and satellite applications, where size, weight, cost, performance, ease of installation, and aerodynamic profiles are constraints, low profile antennas may be required. Presently, there are many government and commercial applications such as mobile, radio, and wireless communications that have similar specifications. To meet these requirements, microstrip antennas can be used [13]. In the literature, artificial neural network (ANN) models have been built for the

Figure.1 CANFIS Architecture Figure.1 shows the adaptive neuro fuzzy inference system. Fuzzy a rule-based systems and artificial neural networks originated from different philosophies and were originally considered independent of each other. Later studies revealed that

they actually have a close relation. Buckley et al. [1] discussed the functional equivalence between neural networks and fuzzy expert systems. The integration of fuzzy logic and neural networks has given birth to an emerging technology field, fuzzy neural networks. The theory of fuzzy logic provides a mathematical strength to capture the uncertainties associated with human cognitive processes, such as thinking and reasoning, also it provides a mathematical morphology to emulate certain perceptual and linguistic attributes associated with human cognition. While fuzzy theory provides an inference mechanism under cognitive uncertainty, computational neural networks offer exciting advantages such as learning, adaptation, fault-tolerance, parallelism and generalization. The computational neural networks are capable of coping with computational complexity, nonlinearity and uncertainty. It is interesting to note that fuzzy logic is another powerful tool for modeling uncertainties associated with human cognition, thinking and perception [2,3]. Many authors have proposed various neuro-fuzzy models as well as complex training algorithms. Of these, Jang [4] proposed the famous neuro-fuzzy model ANFIS (adaptive network-based fuzzy inference system), which has beensuccessfully applied in various fields. In ANFIS, the hybrid learning algorithm is adopted that integrates the BP successfully applied in various fields. In ANFIS, the hybrid learning algorithm is adopted that integrates the BP(Backward Propagation) algorithm with the recursive least squares algorithm to adjust parameters. ANFIS was laterextended to the coactive ANFIS in [5] and to the generalized ANFIS in [6]. Horikawa et al. [7] proposed a neuro-fuzzy model using sigmoid functions to generate the bell-shape input membership functions and trained it with the BP algorithm. However, some practical difficulties associated with gradient descent are slow convergence and ineffectiveness at finding a good solution [8]. The Proposed Neuro-Fuzzy Model The proposed neurofuzzy model is a multilayer neural network-based fuzzy system and the system has a total of five layers. In this connectionist structure, the input and output nodes represent the input states and output reponse, respectively, and in the hidden layers, there are nodes functioning as membership functions (MFs) and rules. This eliminates the disadvantage of a normal feedforward multilayer network, which is difficult for an observer to understand or to modify. Throughout the simulation examples presented in this paper, all the MFs used are bell-shaped (Gaussian) functions defined

as: (x) = exp - ((x- c)2 / 2 ) A Gaussian membership function is determined by c and : c represents the centre of the MF; and determines the width of the MF. A detailed description of the components of the model's structure and functionalities, and the philosophy behind this architecture are given below.

a.

Input Layer

Nodes in this layer are input nodes that represent input linguistic variables as crisp values. The nodes in this layer only transmit input values to the next layer, the membership function layer. Each node is connected to only those nodes of layer 2, which represent the linguistic values of corresponding linguistic variables. b. Fuzzy Input Layer

Nodes in this layer act as membership functions to represent the terms of the respective linguistic variables. The input values are fed to fuzzy input layer that calculates the membership degrees. This is implemented using Gaussian membership functions with two parameters, mean (or centre, c ) and variance (or width, ).This layer implements fuzzification for the inputs. It represents fuzzy quantization of input variables. yt( FI) = exp - ( ((x(I ) - c)2 / 2 ) c. Rule Nodes Layer

The third layer contains rule nodes that evolve through learning. Evolving means all nodes on the third layer are created during learning. The rule nodes represent prototypes of input-output data associations that can be graphically represented as associations of hyper-spheres from the fuzzy input and the fuzzy output space. Hence, the functions of the layer are yt( R) = min yi( F) ; with I is the element of It where I t is the set of indices of the nodes in fuzzy layer that are connected to node t in Rule layer and yi( F) is the output of node i in Fuzzy input layer. d. Fuzzy Output Layer

The fourth layer is fuzzy output layer where each node represents fuzzy quantisation of the output variables. The activation of the node represents the degree to which this membership function is supported by all fuzzy rules together. The connection weights wkj of the links connecting nodes k in fuzzy output layer to nodes j in rule nodes layer represent conceptually the CFs of the corresponding fuzzy rules when inferring fuzzy output values. yt( R) = min (yi(R) wik(F)) where Ik is the set of indices of the nodes in Rule layer that are connected to the node k in Fuzzy output e. layer. 5. Output Layer

given in Table 1 and. Simulation has done on Neurosolution 6 and CST 4. References

This represents the output variables of the system. These nodes and the links attached to them act as a defuzzifier. A node in this layer computes a crisp output signal. The output variable layer makes the defuzzification for fuzzy output variables. The input-output relationship of the units in each layer are defined by the following equations:

yt

w

k k

FO, R tk

FO, R tk

Where ctk and tk are respectively, the centroid and width of the membership function of the output linguistic value represented by k in Fuzzy output layer. 3. Results A. Structures of the CANFIS The CANFIS network, which has a configuration of 4 input neurons, 10 and 5 neurons in 2 hidden layers, and 2 output neurons with learning rate = 0.1, goal = 0.001, was trained for 400 epochs. The membership function is Gaussian and fuzzy model is tsukamoto. Sigmoid axiom with learning rule momentum is used in training. CANFIS network are shown in Table 1. CANFIS is trained with 45 samples and tested with 21 samples determined according to the definition of the problem; 5 inputs and 2 outputs were used for the analysis of CANFIS. The results of the analysis CANFIS for an isotropic material ( r= y) and comparison with the targets are

[1] Balanis C.A.,1997: Antenna theory, 3rd Edition, John Wiley and Sons, Hoboken, NJ. [2] Bahl I.J. and Bhartia P.,1980: Microstrip antennas, Artech House, Dedham, MA. [3] D.M. Pozar,1992: Microstrip antennas, Proceeding, IEEE 80, 7981. [4] Buckley J.J., Hayashi Y.,1994: Fuzzy neural networks: a survey, Fuzzy Sets and Systems, 1-13 [5] Chokri S., Abdelwahed T, 2003: Neural Network for Modeling: A New Approach. Springer-Verlag Berlin Lecture Note in Computing Science,159-168 [6] Takagi H.,1995:"Fusion techniques of fuzzy systems and neural networks, and fuzzy systems and genetic algorithms, SPIE 2061 pp 402-413. [7] Jang J.-S.,ANFIS: adaptive-network-based fuzzy inference system, IEEE Trans, Cybernet.23 pp 665685 1993. [8] Mizutani Eiji, Jang J.-S., 1995: Coactive neural fuzzy modeling, proceedings,IEEE International Conference on Neural Networks,Vol.2,Perth, Australia, 760 765. [9] Azeem M.F. et al., 1994: Generalization of adaptive neuro-fuzzy inference systems, IEEE Trans.Neural Networks 11 pp 1332 1346 2000. [10 ]Horikawa Shin-ichi,et al., 1992:On fuzzy modeling using fuzzy neural networks with the backpropagation algorithm, IEEE Trans.Neural Networks 3 pp 801 806. [11] G. Puskorius, L. Feldkamp, Neurocontrol of nonlinear dynamical systems with Kalman filter trained recurrent networks, IEEE Trans. Neural Networks 5 pp 279297. [12] L. Zadeh, 1965: Fuzzy Sets, Inf. Control, vol 8 pp 338-353. [13] Mamdani, E. H. and Assilian S., 1975:"An experiment in linguistic synthesis with a fuzzy logic controller." Int. J. Man-Machine Studies 7.

Table.1 Results of the CANFIS algorithm and comparison with the targets. h 2.33E+00 2.33E+00 2.33E+00 2.50E+00 2.40E+00 2.20E+00 1.00E+00 1.40E+00 1.70E+00 1.90E+00 1.00E+00 1.80E+00 1.20E+00 1.70E+00 1.50E+00 1.60E+00 1.70E+00 1.80E+00 1.80E+00 1.70E+00 1.70E+00 1.75E+00 1.20E+00 1.50E+00 1.50E+00 1.60E+00 1.60E+00 2.00E+00 2.20E+00 2.50E+00 2.70E+00 2.90E+00 2.60E+00 3.00E+00 2.40E+00 2.10E+00 2.80E+00 2.90E+00 2.70E+01 2.80E+00 2.90E+00 2.60E+00 f 7.55 8 7.75 8.56 7.15 7.28 7.92 7.32 6.69 6.37 6.4 7.03 6.6 6.7 8.64 6.28 6.08 11 6.81 6.66 7.28 DESIRED (W ) 15.4 14.53 14.99 13.24 16.09 16.28 15.3 15.6 15.9 13.9 13.7 13.2 14.4 15.4 13 15.2 15.4 15.8 14.8 14.6 13 DESIRED (L) 11.95 11.6 11.78 10.21 12.61 12.89 12.6 12.8 13.1 13.3 13.5 12.2 12.4 13.4 11 13.2 13.4 13.8 12 12.2 11.8 CANFIS (W) 15.4148 14.5349 14.986 13.2475 16.1 16.28608 15.3407 15.6392 15.9035 13.8848 13.7933 13.264 14.4188 15.4963 12.9528 15.2198 15.4497 15.7943 14.8078 14.6154 13.08 CANFIS (L) 11.9469 11.6029 11.8052 10.2011 12.6442 12.8399 12.6077 12.8634 13.1433 13.3094 13.5376 12.2472 12.4326 13.4356 10.9998 13.202 13.4602 13.7897 12.0884 12.2632 11.8101

S.No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

- 4_1996-Fuzzy Logic=Computing With WordsUploaded byDr. Ir. R. Didin Kusdian, MT.
- redes neuronalesUploaded byjota de copas
- genetic fuzzy logicUploaded byPriya Modi
- BrainSlug- Transparent Acceleration of Deep Learning Through Depth-First ParallelismUploaded byJexia
- Intelligent Decision Support Systems.pptUploaded byOsmn Kmyo
- bell2Uploaded bySarbjeet Singh
- A PREDICTIVE MODEL FOR MAPPING CRIME USING BIG DATA ANALYTICS.pdfUploaded byesatjournals
- Optimisation of Composite Boat Hull StructuresUploaded bymuseblade
- 2-383-15051090329-11 (1)Uploaded byAshish Tom
- a Neuro-Fuzzy Controller for Collaborative Applications in Robotics Using LabVIEWUploaded byYasser Safan
- Week11 Fuzzy Logic TutorialUploaded byAzwar Genetika
- prl16rrrUploaded bySatyabrat Srikumar
- A Multiple Level MIMO FL Based Intelligence For Multi Agent Robot SystemUploaded byAnonymous vQrJlEN
- Paper 14Uploaded byasifmechengr
- Medical Diagnosis ClassificationUploaded byCS & IT
- 10.1.1.21.7030Uploaded bycaldevilla
- Neutrosophic Logic Approaches Applied to ”RABOT” Real Time ControlUploaded byAnonymous 0U9j6BLllB
- Scheduling Jobs on Parallel Machines Applying NeuralUploaded byAntoniojuarezjuarez
- Determination of Combustion Parameters Using Engine Crankshaft SpeedUploaded bysundari_murali
- Research on Active Control of Diesel Engine Based on Wavelet Neural NetworkUploaded byJean Carlos Pineda
- AL21234240Uploaded byAnonymous 7VPPkWS8O
- Salto hidraulicoUploaded byMiguelito88819
- 10.1109@CDC.1989.70217Uploaded byRahim Ali Qamar
- 3535Uploaded byJosue Marshall
- 1701.04928v1.pdfUploaded byNick Pappas
- 1506.06442v1Uploaded byVedran Karamani
- Scimakelatex.13332.One.two.ThreeUploaded bymdp anon
- 9. Neural Networks - LearningUploaded byoletoro14
- 5-fuzzyUploaded by3Dotzz RnD
- 32bfe5113589d4968aUploaded byjuanomoran

- Matlab, Simulink Building a Direct Sequence Spread Spectrum ModelUploaded bysir_jon43
- RTU Thesis ch-1 introUploaded byAlok Pandey
- chapter-5Uploaded byAlok Pandey
- Chapter 4Uploaded byAlok Pandey
- ResumeUploaded byAlok Pandey
- Alok ResumeUploaded byAlok Pandey
- Ch-2 Literature SurveyUploaded byAlok Pandey
- Chapter 5 Conclusion & Future WorkUploaded byAlok Pandey
- Est Solution 4Uploaded byAlok Pandey
- Digital signal processing solution mannual chapter 02Uploaded byXafran Khan
- Chapter 3Uploaded byAlok Pandey
- AbstractUploaded byAlok Pandey
- Rlative Motion Very ImpUploaded byAlok Pandey
- IntroductionUploaded byHizbullah
- Class 11 PhysicsUploaded byAlok Pandey
- Adaptive Antennas at the Mobile Handset for Reduced Radiation Hazards an AnalysisUploaded byAlok Pandey
- GM PU Supervisor-FormUploaded byAlok Pandey
- 3uf_phdadmissionformUploaded byAlok Pandey
- Cameraready PaperUploaded byAlok Pandey
- 10.08120101Uploaded byAlok Pandey
- rls algorithmaUploaded byAlok Pandey
- decomposition methodUploaded byAlok Pandey
- New Microsoft Word DocumentUploaded byAlok Pandey
- seminar guidelinesUploaded byAlok Pandey

- An Artificial Intelligence Based Neuro Fuzzy SymUploaded bybvinothhh
- 1 ANFIS Based Neuro-Fuzzy Controller in LFC of Wind-Micro Hydro-Diesel Hybrid Power SystemUploaded bysatti849
- ADAPTIVE NETWORK BASED FUZZY INFERENCE SYSTEM FOR SPEECH RECOGNITION THROUGH SUBTRACTIVE CLUSTERINGUploaded byAdam Hansen
- NeuroUploaded bydennyyusuf
- Adaptive & Optimal Control SystemsUploaded byDarryl Shelton
- 2012-1161. Neuro-Fuzzy Systems.pptUploaded bylutfijuhar
- A novel power swing blocking scheme.pdfUploaded byRaviYadav
- Fuzzy SetsUploaded byaryo_el06
- Simpliﬁed Fuzzy Rule-based Systems using Non-parametric Antecedents and relative Data Density.pdfUploaded byOualid Lamraoui
- “Prediction of Industrial Solid Waste with ANFIS Model and its comparison with ANN Model- A Case Study of Durg-Bhilai Twin City India,.pdfUploaded byMk Tiwari
- Mang AoUploaded byvantienbk
- Ch 7 Artificial Intelligence Techniques.pdfUploaded bymaheshnagarkar
- B2Uploaded byvishwaharnal115
- Online-Adaptif-Neuro-Fuzzy-Control.pptUploaded byArioston Júnior
- 10.1.1.42Uploaded byaharisjo
- International Journal of Advance & Innovative Research Volume 2, Issue 2 ( II ) April to June 2015 ISSN: 2394-7780Uploaded byempyreal
- a Neuro-Fuzzy Controller for Collaborative Applications in Robotics Using LabVIEWUploaded byYasser Safan
- Jp St 4451425155400Uploaded byms_petro83
- Vehicle License Plate Recognition SystemUploaded bydips032
- Generating fuzzy networksUploaded bylenomagnus
- 0456_PDF_09Uploaded byLUSS4U
- anifowose2016Uploaded byNguyenHuy
- Contents Artificial IntelligenceUploaded byJulio Anthony Leonard
- ( ) D. Dubois, H. Prade (auth.), M. J. Patyra, D. M. Mlynek (eds.)-Fuzzy Logic_ Implementation and Applications-Vieweg+Teubner Verlag (1996)Uploaded bybathaie_mohammad
- Intelligent-Temperature-Controller-for-Water-Bath-System.pdfUploaded byOmprakash Verma
- Adaptive Neuro-Fuzzy Controller for Robot VehicleUploaded byanaffira
- Adaptive Neuro FuzzyUploaded byseembab
- Prioritization of Detected Intrusion in Biometric Template Storage for Prevention using Neuro-Fuzzy ApproachUploaded byJournal of Computing
- (Advances in Information Security 70) Ali Dehghantanha, Mauro Conti, Tooska Dargahi-Cyber Threat Intelligence-Springer International Publishing (2018).pdfUploaded byirischan
- A Neuro-fuzzy Multiple-model Observer Approach to Robust Fault Diagnosis Based on the DAMADICS Benchmark ProblemUploaded byJuan David Bolaños Aguilar