You are on page 1of 7

Ocean Engineering 131 (2017) 224–230

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Design and performance analysis of a tracked wall-climbing robot for ship MARK
inspection in shipbuilding

Haocai Huanga, Danhua Lia, Zhao Xueb, XianLei Chenc, Shuyu Liua, Jianxing Lenga, Yan Weia,
a
Ocean College, Zhejiang University, Zhoushan, 316021 Zhejiang, China
b
Development Center of Qingdao Nation Laboratory for Marine Science and Technology, Qingdao, 266237 Shandong, China
c
Zhoushan Institute of Calibration and Testing for Quality and Technology Supervision, 316021 Zhoushan, China

A R T I C L E I N F O A BS T RAC T

Keywords: Ship inspection is renowned for its hazardous and cluttered working environment. In worst cases, human
Ship inspection operators must risk their lives to reach the almost inaccessible inspecting area. This situation yields, in a natural
Wall-climbing robot way, to the utilization of remotely controlled devices. Thus a tracked wall-climbing robot was brought forth in
Probe clamping device our research as one of the most advanced and promising solutions. In this paper, a robot with climbing ability is
Magnetic track
designed and prototyped for vertical structure inspection in shipbuilding. It combines magnetic caterpillars and
magnets to climb ferromagnetic surfaces, like double-bottom of ship hull. The structure of two trapezoidal
tracks allows the robot to climb over uneven surfaces, like ship bulwark. Equipped with special probe clamping
devices and employing magnetic properties, the robot is able to perform ship inspection tasks in those human
unfriendly and inaccessible environments and therefore cut down costs. This achievement may open new
avenues for mobile robotics inspection of ferromagnetic industrial structures as well.

1. Introduction adhesion mechanism, five categories are classified, namely adhesion


using suction and propulsion, magnetic adhesion, adhesion using
The regular inspection of ship ensures the operational safety and gripping equipment, biomimetic adhesion (Lynch et al., 2012) and
prevents pollution to marine environment. It provides necessary data rail-guided adhesion (Chu et al., 2010; Schmidt and Berns, 2013; Shen
not only for government to handle the management and control et al., 2005). In real applications, such as Neptune (Schempf et al.,
problems, but also for the ship owners to improve competitiveness in 1995), Tripillar (Schoeneich et al., 2011), Combot (Lee et al., 2012) and
the market. The major inspection and maintenance tasks protect the HydroCrawler (JetEdge, 2010), combined types of locomotion and
ship hull from unexpected leaks which may occur by cracks and pitting adhesion mechanisms could be selected in order to meet specific
caused by corrosion and weld seams (Kalra et al., 2006). However, ship requirements.
inspection is also renowned for its hazardous and cluttered working The earliest wall climbing robots can be dated back to the 1960s,
environment. In worst cases, inspectors must risk their lives to reach developed by Nishi (Nishi et al., 1986) and were able to move on
the almost inaccessible inspecting area. To reduce the heavy manual vertical walls and ceilings but with the shortcomings of large size and
labor and eliminate the inspection work in a harsh environment, the heavy weight. Since then, most researchers have been focusing on
shipbuilding industry has shown a great potential for the applications designing climbing robots operating on vertical human-made struc-
of wall-climbing robots. tures. And they developed the first prototype in the 1990s (Schmidt and
Being a kind of special mobile robot, wall-climbing robot is often Berns, 2013). It demonstrated a great number of potential applications
designed to work on the vertical plane. It is required not only to have that cannot be performed directly by human operators in a proper and
the same locomotion mechanism as traditional mobile robots, but also safe manner. The application fields vary from inspection of nuclear
to possess the unique characteristics of sustaining itself against gravity power plants, cleaning of high-rise buildings to welding and main-
while moving (Balaguer and Armada, 2006; Prahlad et al., 2008; Tong tenance of ship hulls. For such tasks, often-cited Systems REST (Grieco
et al., 2014). In accordance with the locomotion type, the climbing et al., 1998), ROBUGⅡ(Luk et al., 1991) or NINJA (Nagakubo and
robot can be mainly divided into four classes, which are arms and legs, Hirose, 1994) were designed. Although many applications can be
wheels and tracks, wires and rails and sliding frame. With regard to the realized by means of appropriate manipulator equipped with the right


Correspondence author.
E-mail address: redwine447@zju.edu.cn (Y. Wei).

http://dx.doi.org/10.1016/j.oceaneng.2017.01.003
Received 18 October 2016; Received in revised form 26 December 2016; Accepted 3 January 2017
0029-8018/ © 2017 Elsevier Ltd. All rights reserved.
H. Huang et al. Ocean Engineering 131 (2017) 224–230

tools plus the support of the human operator's skills, many others
cannot be solved in this way, e.g. ship inspection, one of the most
urgent challenges due to the difficult access to working conditions
(Armada et al., 2005; Balaguer, 2005). Our proposed wall- climbing
robot is to carry NDT (non-destructive testing) tools for ship inspec-
tion. Through the remote control, the robot is capable of moving freely
within a confined space. We choose tracked mechanism and permanent
magnetic adhesion type with properties of high reliability, simple
controlling method, and high speed.
This paper is organized as follows: Section 2 presents an overview
of the robotic system and describes the characteristics in order to
accomplish the ship inspection tasks in an optimal way; the design of
robot including the mechanical architecture, the detection mechanism
and the control system are detailed in Section 3; in Section 4, the force
analysis for safety is instructed; Section 5 analyzes its self-adjusting
performance and experiments and discussion are described in Section
6; the conclusion ends the paper.

2. System overview

The wall-climbing robot in this paper targets on ship inspection, so


other excessive functions of general robots are excluded. The robot
consists of four modules including the magnetic track system, control
Fig. 1. Mechanical structure (1 ruler, 2 the small sprocket, 3 front support bracket, 4
system, holder for detector and the robot body structure. Since ship
driven wheel, 5 small axle, 6 driven axle, 7 screw base, 8 lead screw, 9 micro motor, 10
hull mostly made of steel is often uneven due to welding deformation, baseboards, 11 crawling chain and magnet, 12 tensioning sprocket, 13 back support
the permanent magnetic is chosen as the adhesion and locomotion bracket, 14 driving wheel, 15 chain of transmission, 16 driving axle, 17 motor and
type. Its track is driven by a motor. Since a special holder is designed to reducer, 18 axle bed).
carry different detecting instruments and allows the robot to perform
ship inspection in various occasions, its payload capacity is increased others are used for tensioning the chain. In order to adjust the
accordingly. sprocket, a tank is designed to sit on the rear support bracket. Due
The wall-climbing robot has the following characteristics: to the existence of welding joints on the surface of ship hull, an angle of
approximately 45° is inclined between the driving sprocket and the
(1) A wired remote control is applied, which is able to set the multi- tensioning sprocket to enable the robot to climb over the uneven
pole mobile speed by pressing different keys. This helps to reduce surface. Fig. 2 is the prototype picture of the robot from the front view
the work intensity of manual measurement and further improves and the side view respectively.
the inspection accuracy.
(2) A length adjustment mechanism is designed and the length of
crawl can be changed appropriately in accordance with the
requirement.
(3) The payload capacity is strong. Either track is installed with 66 3.2. Inspection system
permanent magnetic units. All of them are fixed by an aluminum
plate to avoid the rust suction. The climbing robot is developed to achieve the goal of automatic
(4) Small size and light weight with large adhesion force make the detection of ship hull. The ability of carrying various detection
robot portable and stable. It is convenient for operators to navigate instruments is rather crucial, which enables the robot to be applied
in the field. to different occasions. Therefore, a probe clamping device is designed
and attached to the robot. The probe clamping device is shown in
3. Design of a caterpillar wall-climbing robot Fig. 3.
Here, we choose the ultrasonic detection probe in the field of non-
3.1. Mechanical system destructive testing since it is a mature product with lower possibility of
being disturbed by other factors. The ultrasonic flaw detecting module
The mechanical structure of the climbing robot is shown in Fig. 1. It consists of ultrasonic detector and its driven software, the ultrasound
has the dimensions of 300 mm×200 mm×83.54 mm with the weight of probe and its clamping device. Its working method is explained as
10 kg. As the climbing robot is specialized in inspection, it consists of follows:
permanent magnetic adhesion mechanism, tracked locomotion me-
chanism, driving mechanism, the frame of the body and inspection 1. The probe clamping device lifts or drops the probe with the
system capable of carrying out detection missions. Either track is operation of worm and gear reducer driven by the control system.
comprised of a roller chain inside of which are 66 arranged permanent 2. The probe clamping device holds tightly and locates accurately the
magnetic units. An alternating current (AC) servo motor is chosen to be probe via the big spring, the small spring, the driving wheel and the
the robot's power producer due to its small in size, lightness in weight hinge.
and being easily controlled by microcomputer. The motor power is
30 W and uses 220 voltage alternating current. Through a reducer with Equipped with the probe clamping device, the robot is able to take
the ratio of 1:36, the motor shown part 17 in Fig. 1 delivers the power the ultrasonic probe to the detection area which is too risky for the
to the driving sprocket through the chain, and the sprocket drives the manual detection. To accomplish the task, the movement of robot with
main axle as well as the magnetic track. Then the magnetic track the probe clamping device follows the command of control system
receives power for going forward. Inside the track are four sprockets, which will be described in the next section Fig. 4 shows the monitoring
two of which are utilized to pass power to the roller chain and the device.

225
H. Huang et al. Ocean Engineering 131 (2017) 224–230

Fig. 2. The front view and side view of the prototype.

Fig. 5. The wired remote control of thickness measurement.


Fig. 3. A probe clamping device (1 worm gear, 2 worm, 3 small spring, 4 the water inlet
hole, 5 rolling wheel, 6 nut, 7 screw, 8 big spring, 9 hinge, 10 probe). moving speed can be set by pressing different functional keys. This
mode allows the robot to be controlled by an operator on the work field
without any delay; it can also be controlled by the industrial computer
via communication system. Fig. 5 is an example of the wired remote
control mode of thickness measurement.
The control system of the climbing robot is based on hierarchical
control architecture (Schmidt and Berns, 2013). It includes the host
computer and guest computer as suggest by Yi (Yi et al., 2009). The aim
of the control hierarchy is to relieve the robot operator from arduous
low-level control through converting a series of high-level tasks into a
sequence of driving signals.
An industrial control computer (ICP) is used here as a host
computer. It is responsible for the management and control of the
entire working network and then sending orders to the guest computer,
realizing the remote control of the robot. Its function includes
Fig. 4. Monitor for ship inspection.
preliminarily determining the walk path, displaying its work condition,
implementing system's self-diagnosis/protection etc. We designed
3.3. Control system
different function keys on the host computer's circuit board, represent-
ing different rotating speed of servo motor and thus corresponding to
The control system of the climbing robot which controls its move-
various kinds of movements of robot respectively.
ments and navigates it along a predetermined trajectory is a very
Guest computer system, as the core of the entire control system,
significant part. It has the following characteristics:
uses single chip microcomputer. The whole set controller, installed at


the back of the robot body, consists of an AC servo motor and AC servo
reliable, convenient and flexible to operate.

driver, a DC power supply module and a control circuit board. The
In real time the moving directions and moving speed can be
control circuit board is made up of a AT89C52 single chip micro-
adjusted.

computer, a 8155 extended I/O interface, an operational amplifier
a wired remote control is applied.
circuit, an A/D transform circuit, a digital I/O interface circuit and a
standard RS-485 interface. The main function of the guest computer
Due to the harsh working environment, the control mode is limited
system is to receive and process the orders from the host computer and
and here, a wired remote control mode is chosen via which the multiple
then output the analog voltage signals to control servo drivers which

226
H. Huang et al. Ocean Engineering 131 (2017) 224–230

Fig. 6. Configuration of the control system of the wall-climbing robot.

further control the rotation of the servo motor and the inspection
instruments so as to control robot motions. The configuration of the
control system of the wall-climbing robot is illustrated in Fig. 6.

4. Force analysis

In order to improve the robot motion performance, we need to deal


with several potential risks which are, firstly, rolling around the last
contact magnetic point between the track and the wall, secondly,
sliding along the wall, and thirdly, deviating from the routine (Gao
et al., 2009; Yi et al., 2009). In any of these cases, the robot will not
accomplish the inspection task successfully. To ensure the stability of
operation and safety of adhesion, we need to increase the magnetic
Fig. 7. Static force analysis of the robot.
adhesion force, while it would inevitably lead to the reduction of the
flexibility. So it is necessary to build a force analysis model to balance
balance status, normal to the sliding surface direction we have (Nishi
them two.
et al., 1986; Yi et al., 2010),
When the robot is resting on the vertical surface, it bears forces
n
such as the gravity, the supporting force of the surface, the adhesion
force, and the friction between the tracks and the surface (Schmidt and n⋅Fm = ∑ Ni
i =1 (1)
Berns, 2013). The force analysis is done on random pose of the robot
operating on the wall at any inclination angle. The studies under where Ni is the supporting force of the surface to each permanent
different situations result in a general analysis method about static magnetic unit, n is the number of permanent magnetic units adsorbing
force equilibrium. A robot climbing model is built in Fig. 7 (Jiang et al., on the surface and Fm is the adhesion force of each permanent magnetic
2010). unit.
Our aim is to make sure that the climbing robot is able to adhere to Every magnetic unit on the crawler bears the same adhering force,
a surface in an arbitrary angle. In Fig. 7, C is the crawler's last contact but to each magnetic unit, the supporting force of the surface Ni caused
point with the surface, Fm1 is the first permanent magnetic unit by the torque of gravity differs.
adhesion force normal to the sliding surface, Ff is the sliding friction As shown in Fig. 7, to prevent the robot being overturned around
force, G is the gravity plus the loading weight of the robot, l1 is the the point C, the adhesion force of magnetic units must overcome the
distance between the bottom magnetic unit and the crawler gravity total torque which arises from the gravity. Since the chain in the track
center along the sliding surface, l is the distance of the crawler adhering is flexibly connected by numbers of small joints, the torque at point C
to the surface, h is the perpendicular distance between gravity center cannot be transmitted by the chain. So, if the robot rotates around the
and the wall and α is the inclination angle of the sliding surface. point C, the first upper magnetic unit will be lifted. At point C, the
If the robot rotates around the point C, while it still maintains the torque balance equation is (Nishi et al., 1986; Yi et al., 2010)

227
H. Huang et al. Ocean Engineering 131 (2017) 224–230

∑ Mc = M1 + M2 + M3 = 0 (2)
where M1 is the torque of the first magnetic unit, M2 is half of normal
torque of gravity, and M3 is half of tangent torque of gravity. We have,
M1 = (Fm1 − N1) l
(G ⋅ sin α ) l1
M2 = 2
Gh ⋅ cos α
M3 = 2 (3)
If the robot overturns and rotates around the point C, then,
(G ⋅ sin α ) l1 Gh ⋅ cos α
(Fm1 − N1) l − 2
− 2
=0
(G ⋅ sin α ) l1 Gh ⋅ cos α
Fm1 = 2l
− 2l
+ N1 (4)
If the robot does not overturn, the first upper magnetic unit should
be able to adhere to the surface, then,
N1 > 0 (5)
According to (4) and (5), we have,
(G⋅ sin α ) l1 Gh⋅ cos α Fig. 9. A supplemental correction part.
Fm1 ≥ −
2l 2l (6)
α, β, γ are three angles of the climbing robot to the ideal wall
Therefore, each magnetic unit should fulfill (6) so that the robot will
surface (Kalra et al., 2006). In fact, it is assumed that γ and α are equal,
not overturn from the wall during operation.
thus only two position sensors are needed for the offset correction. If
the offset d appears obviously caused by deformation of wall, the
5. Self-adjusting analysis
control system can move one track faster to ensure the robot to climb
vertically. The angular deviation of the robot in this paper is less than
It is known that the surface of ship is not flat due to welding seams,
1°50′, referred to the vertical line.
deformation of wall and corrosion of the material. Therefore, the pure
design of a flexible track to adapt to the deformation is not sufficient.
The robot may not follow the predetermined path to walk. Examples of 6. Experimental tests
the climbing robot resting on uneven surfaces are shown in Fig. 8.
To solve the mentioned problem above and further improve the When the climbing robot with the detection instruments conducts
accuracy as well, a supplemental correction part is designed (Chen inspection work, it may fail to complete the missions and even slides
et al., 1997) and shown in Fig. 9. It makes the robot self-adapt to down or overturns in extreme cases, if its payload is overloaded or the
uneven surfaces. Meanwhile, the self-correction function is added to operation parameters are not properly set. To verify the performance
the control system. Whenever there is a difference between the actual and validate the parameters of the robot, a series of experimental tests
trajectory and the designed trajectory, the offset value could be were conducted in the laboratory. The tests include climbing speed
detected and corrected. The correction process is described as follows. testing, payload capacity testing, and obstacle-overcoming perfor-
In Fig. 10, the vertical height h that the robot has to climb over can mance testing.
be expressed as:
6.1. Climbing speed test
h = KL cos γ (7)
where L is the real length of the trajectory, K is a proportionality factor, It is a fact that the climbing speed would influence the detection
and γ is the angle between the robot and the vertical axis. efficiency. Therefore the speed was tested and described as follow:
The ruler carried by the robot as shown in Fig. 9 is used to measure
the angle γ , using horizontal projection. We have, (1) Test conditions. The climbing distance was measured by a steel
(8) tape and a stopwatch was used to record the climbing time. The
r = R cos α
maximum permissible error range of the tape was ± (0.1+0.1 L)
where R is the dimension measured by the ruler,α is the angle between mm and the range of the stopwatch is ± (T×7×10-7+3 ms),
the ruler and the horizontal axis and r is the horizontal projection of R . where L was the climbing distance and T was the climbing time.
If the robot deviates from the designed path, the offset that should (2) Test method. The robot moved on a lab table and its travel distance
be corrected is and time were recorded respectively. Thus the moving speed could
d = b tan β (9) be obtained. The robot was tested when its motor ran in four kinds
of frequency, which were 30 HZ, 40 HZ, 50 HZ, and 60 Hz. For
where d is the offset to be adjusted, b is the body width, and β is the each frequency, the speed was tested three times. Here, the
angle from the vertical line. walking distance was set to 1.8 m. The test result is shown in
Table 1.

(3) Test conclusion. The robot has a wide range of speed which mainly
depends on the frequency of motor. The maximum speed is even
over 7 m/min.

6.2. Adsorption capacity test

In general, three types of coating are predominantly used in ship


Fig. 8. Examples of the climbing robot on uneven surfaces. hull painting, which are epoxy, polyurethane and alkyd. These layers

228
H. Huang et al. Ocean Engineering 131 (2017) 224–230

Fig. 10. Analysis of the deviation correction.

Table 1 be 0 mm, 0.5 mm, and 1 mm respectively.


Climbing speed test result. (2) Test method. During the test, the robot was placed on iron plates
with different coating layers. The absorption force was tested by a
Frequency No. Distance Time Speed
(m) (s) (m/min) dynamometer. The test result is demonstrated in Table 2.

30 Hz 1 1.8 39.2 2.76 (3) Test conclusion. With the increase of coating thickness, the
2 1.8 42 2.57 adsorption force reduces. When coating thickness reaches 1 mm,
3 1.8 41.3 2.62
average 1.8 40.8 2.65
the robot is still able to adsorb on the ship surface firmly. Since for
common steel plates, the thickness of coating is between 50 µm
40 Hz 1 1.8 21 5.14 and 320 µm, it is believed that the maximum test thickness set to
2 1.8 20.8 5.19 be 1 mm is sufficient to prove the robot's adsorption capacity.
3 1.8 20.1 5.37
average 1.8 20.6 5.23
6.3. Payload capacity test
50 Hz 1 1.8 17.3 6.24
2 1.8 16.7 6.47 (1) Test condition. The robot carried different weights to climb the
3 1.8 17.1 6.32 surface with 0.5 mm thickness of painting layer. In Fig. 11, the
average 1.8 17.0 6.34
robot born standard weights, i.e., 1 kg, 2 kg and 3 kg.
60 Hz 1 1.8 14.7 7.35 (2) Test method. When the robot was climbing on the wall in the lab,
2 1.8 15.4 7.01 increase the weights until it reached the critical state in which the
3 1.8 14.9 7.25 robot had a tendency of slipping down. Calculate the total weight
average 1.8 15.0 7.20
and then repeat the above steps.
(3) Test conclusion. When the weights was increased to 6 kg, the robot
cannot climb. So its payload capacity should be less than 6 kg.
Table 2
Adsorption capacity test result.
6.4. Obstacle-overcoming performance test
Coating thickness (mm) 0 0.5 1
Adsorption force(N) 1705 1056 667
(1) Test condition. Several iron plates with the thickness of 10 mm
were used.
(2) Test method. The iron plates were placed in different locations
are sprayed evenly on the ship surface, which directly affect the robot's
along the moving trajectory of the robot, as shown in Fig. 12. The
absorption force. Thus its adsorption capacity needs to be tested as
robot moving stability was observed.
well.
(3) Test conclusion. The robot could overcome obstacles with certain
height, which was no less than 10 mm.
(1) Test conditions. We selected the same non-magnetic painting used
on ship surface, i.e., epoxy paint. The coating thickness was set to

Fig. 11. Payload capacity test process with 1 kg, 2 kg and 3 kg weights respectively.

229
H. Huang et al. Ocean Engineering 131 (2017) 224–230

Balaguer, C., Armada, M., 2006. Robot applications against gravity. IEEE Robot. Autom.
Mag. 13 (1), 5–6.
Balaguer, C., Gimenez, A., Jardon, A., 2005. Climbing robots' mobility for inspection and
maintenance of 3D complex environments. Auton. Robots 18 (2), 157–169.
Chen, J., Cheng, J., Feng, P., Ma, P., 1997. Research of control system of climbing robot
for oil tank volume measurement. Ind. Instrum. Autom., 7–12.
Chu, B., Jung, K., Han, C.S., Hong, D., 2010. A survey of climbing robots: locomotion and
adhesion. Int. J. Precis. Eng. Manuf. 11 (4), 633–647.
Gao, X.S., Xu, D.G., Wang, Y., Pan, H.H., Shen, W.M., 2009. Multifunctional robot to
maintain boiler water-cooling tubes. Robotica 27 (6), 941–948.
Grieco, J.C., Prieto, M., Armada, M., 1998. A six-legged climbing robot for high payloads.
In: Proceedings of the IEEE International Conference on Control Applications,
Trieste, Italy, vol.1, pp. 446–450.
JetEdge, 2010. Jet EdgeWaterjets supply 3800 bar Water for your toughest Surface Prep
Projects, Tech. Rep., Jet Edge, St. Michael, MN, USA.
Jiang, H., Yang, X., Jiang, H., 2010. A design and implementation of adsorptive crawl
and speeling robot for detecting and repairing on vessel. Ship Sci. Technol. 32 (9),
31–34, (in Chinese).
Kalra, L.P., Shen, W., Gu, J., 2006. A wall climbing robotic system for non destructive
inspection of above ground tanks. In: Proceedings of the IEEE Canadian Conference
on Electrical and Computer Engineering (CCECE’2006), pp. 402–405.
Fig. 12. Obstacle performance test process. Lee, G., Wu, G., Sun, H.K., Kim, J., Seo, T., 2012. Combot: Compliant climbing robotic
platform with transitioning capability and payload capacity. In: Proceedings of the
IEEE International Conference on Robotics and Automation, pp. 2737–2742.
7. Conclusion
Luk, B.L., Collie, A.A., Billingsley, J., 1991. Robug II: An intelligent wall climbing robot.
In: Proceedings of the IEEE International Conference on Robotics and Automation,
This paper has described a newly designed tracked wall-climbing vol.3, pp. 2342–2347.
Lynch, G.A., Clark, J.E., Lin, P.C., Koditschek, D.E., 2012. A bioinspired dynamical
robot special for non-destructive ship inspection. The robot has a wide
vertical climbing robot. Int. J. Robot. Res. 31 (8), 974–996.
climbing speed range with the maximum speed up to 7 m/min and is Nagakubo, A., Hirose, S., 1994. Walking and running of the quadruped wall-climbing
able to move steadily. It has high payload capacity and excellent robot. In: Proceedings of the IEEE International Conference on Robotics and
obstacle-overcoming performance. With its semiautomatic control Automation, vol.2, pp. 1005–1012.
Nishi, A., Wakasugi, Y., Watanabe, K., 1986. Design of a robot capable of moving on a
system and probe clamping device suitable for different detector, this vertical wall. Adv. Robot. 1 (1), 33–45.
integrated multitask robot could meet various ship inspection require- Prahlad, H., Pelrine, R., Stanford, S., Marlow, J., Kornbluh, R., 2008. Electroadhesive
ments, such as online monitoring, thickness measurement and rust robot-wall climbing robots enabled by a novel, robust, and electrically controllable
adhesion technology. In: Proceedings of the IEEE International conference on
cleaning. Robotics and Automation, Pasadena, USA, pp. 3028–3033.
Schempf, H., Chemel, B., Everett, N., 1995. Above-ground storage tank inspection robot
Acknowledgements system. IEEE Robot. Autom. Mag. 2 (2), 9–15.
Schmidt, D., Berns, K., 2013. Climbing robots for maintenance and inspections of vertical
structures—A survey of design aspects and technologies. Robot. Auton. Syst. 61 (12),
The work was financially supported by the National Natural Science 1288–1305.
Foundation of China (No.41576031), and the Fundamental Research Schoeneich, P., Rochat, F., Nguyen, O.T.D., Moser, R., Mondada, F., 2011. TRIPILLAR: a
miniature magnetic caterpillar climbing robot with plane transition ability. Robotica
Funds for the Central Universities (No. 2016FZA4022). The authors
29 (7), 1075–1081.
would like to thank Zhejiang University in China for providing Shen, W., Gu J., Shen, Y., 2005. Proposed wall climbing robot with permanent magnetic
laboratory space. The valuable technical support from the staff and tracks for inspecting oil tanks. In: Proceedings of the IEEE International Conference
on Mechatronics and Automation, vol. 4, pp. 2072–2077.
students throughout the project is also appreciated. The authors are
Tong, W.J., Chin, C.S., Lin, W.P., 2014. Design of a miniature tri-axis vertical wall
grateful to reviewers for their valuable comments, which improved the structure climbing robot using electroadhesion. In: Proceedings of the IEEE
readability of this paper significantly. Conference on Industrial Electronics and Applications, Hangzhou, China, pp. 473–
477.
Yi, Z.Y., Gong, Y.J., Wang, Z.W., Wang, X.R., 2009. Development of a wall climbing robot
References for ship rust removal. In: Proceedings of the IEEE International Conference on
Mechatronics and Automation, Changchun, China, pp. 4610–4615.
Armada, M., Prieto, M., Akinfiev, T., Fernandez, R., González, P., Garcia, E., Montes, H., Yi, Z.Y., Gong, Y.J., Wang, Z.W., Wang, X.R., 2010. Wall-attachment model and its
Nabulsi, S., Ponticelli, R., Sarria, J., Estremera, J., Ros, S., Grieco, J., Gernandez, G., simulation on a new wall-climbing robot for rust removal. J. Sichuan Univ. (Eng. Sci.
2005. On the design and development of climbing and walking robots for the Ed.) 43 (2), 211–216.
maritime industries. J. Marit. Res. 2 (1), 9–32.

230

You might also like