Wavelet neural networks robust control of farm transmission line deicing robot manipulators
Highlights
► Our results are based on the wavelet neural networks robust control method. ► Wavelet neural network is used to approximate the uncertainties and disturbance. ► Sufficient conditions are provided which guarantee the convergence of the system. ► Experimental results show the superior performance of the proposed method. ► The model principle and method are all based on the real FTLDRM system.
Introduction
It is well known that the nonlinearity of robot dynamics makes the analysis even more complex and difficult than control of linear dynamic systems. Until now, the research for control problem of multi-joint manipulator has attracted great attention, and numerous research results have been appeared [1], [2], [3]. Jin et al. [1] present a robust compliant-motion-control technique for a robot manipulator with nonlinear friction. Chen [2] proposes a dynamic structure neural-fuzzy network via a robust adaptive sliding-mode approach to address trajectory-tracking control of n-link robot manipulator.
As is well known, a large number of power line icing will greatly increase the tension loads of farm transmission line towers [4], [5], [6]. Under serious circumstances, it will lead to collapse of line towers and undermine the safe operation of farm transmission lines. Therefore, it is important to pay attention to the farm transmission line deicing online. Recently, Wang et al. make a micromesh study of the farm transmission line deicing robot manipulators (FTLDRM) and make a marked progress [7]. The farm FTLDRM is a mobile visual dual-arm articulated robot (see Fig. 2), which owns the ability of quick-clear of ice on farm transmission line. In this paper, based on the Lyapunov method and the neural network approximation theory, we detailedly analyze the WNN robust control (WNNRC) of the FTLDRM with uncertainties and disturbances. As far as we know, our results are totally original and novel. The reasons are as follows: firstly, till now, there is no result for the vision-based dynamical analysis of the farm FTLDRM system, and the model principle and method in our paper are all based on the real FTLDRM system, which possess the merits of authenticity and practicability; secondly, the proposed WNNRC scheme possesses superior control performance compared with existing results, which is illustrated detailedly both in simulation and experiment analysis in our paper (Fig. 1).
The paper is organized as follows. Section 2 addresses some preliminaries. In Section 3, firstly, tracking performance of the robust control with known parameters of the controlled plant are presented, after that, the structure and properties of the WNN-based robust adaptive control strategy is derived for achieving satisfied tracking performance. Simulation and experiment are provided to present the exceptional performance of the proposed control method in 4 Computer simulation, 5 Experiment results respectively. Conclusion is finally given in Section 6.
Section snippets
System modeling description of FTLDRM
The simplified planar construction of vision-based farm FTLDRM is shown in Fig. 2, which is equipped with two arms. For the purpose of simplification, we just give the parameters of arm one. The notations in Fig. 2 are listed as follows: {O,X,Y} is the world coordinate system, qi (i = 1,2) denote the ith link angular position (rad), mi (i = 1,2) are links masses (kg), li (i = 1,2) represent links lengths (m), Ii (i = 1,2) are inertia moment of link i about corresponding center of gravity (kg m2).
Mathematical notations and definitions
The
Design of neural network robust controller
To effectively control the multi-arm FHMM system, the WNN-based robust control strategy is proposed here. The block diagram of the WNN-based control system is shown in Fig. 4.
Given a desired trajectory qd(t) ∈ Rn, the tracking errors areand the instantaneous performance measure is defined aswhere Λ ∈ Rn × n is the constant-gain matrix or critic (not necessarily symmetric).
The robot dynamics (1) may be written aswhere the robot
Computer simulation
In this section, the proposed wavelet neural network-based robust adaptive control strategy is applied to control robotic systems to verify its effectiveness. Robotic systems have been known to exhibit complex dynamical behaviors. Several control techniques have been proposed for the robotic systems [1], [2], [3]. It should be emphasized that the development of the WNNRC does not need to know the dynamics of the controlled system. Considering a two-link planar elbow manipulator. Set the
Experiment results
In this section, an experimental setup and system realization are described. Then experimental data is discussed and analyzed. The study utilized the PUMA 560 manipulator in our lab of intelligent automation technology. The sampling time used in the experiments is 11 ms.
The dimensions of the robot are given in Table 2, where subindexes 1 and 2 stand for the first and second link, respectively, and LC stands for the center of mass, Par means parameters, I stands for inertia, and ∆ defines joint
Conclusions
In this paper, an WNN-based robust adaptive control method for farm transmission line deicing robot manipulators (FTLDRM) has been presented. The bounds of the uncertainties are not necessarily known. An WNN system is used to approach the unknown controlled system. It is shown that the proposed control scheme can guarantee estimation convergence and stability robustness of the closed-loop system. As demonstrated in the illustrated simulation and experiment examples, the control scheme proposed
Acknowledgment
This work was supported by National Natural Science Foundation of China (60835004, 41007601), Postdoctoral Foundation of China (20110491241), Foundation of Hunan Province Science and Technology Department (2011RS4035), Key Laboratory Foundation for power technology of renewable energy sources of Hunan Province (2011KFJJ004) and Research Foundation of Hunan Provincial Education Department (10C0356).
Shuning Wei received M.S. degrees in College of Electrical Engineering, Guangxi University, in 2005. She is currently working toward the Ph.D. degree in the College of Electrical and Information Engineering, Hunan University, Changsha, China. Her research interests include robot control technology, neural network control and machine learning.
References (17)
Tracking control basedon neural network strategy for robot manipulator
Neurocomputing
(2003)- et al.
Robust compliant motion control of robot with nonlinear friction using time-delay estimation
IEEE Transactions on Industrial Electronics
(2008) Dynamic structure neural-fuzzy networks for robust adaptive control of robot manipulators
IEEE Transactions on Industrial Electronics
(2008)- et al.
Optimal control design for robust fuzzy friction compensation in a robot joint
IEEE Transactions on Industrial Electronics
(2009) - et al.
Investigation and analysis of ice and snow disaster suffered by Hunan power grid in 2008
Electricity
(2008) - et al.
Current situation of natural disaster in electric power system and counter measures
Journal of Natural Disasters
(2006) - et al.
Review on the wide area blackout of 500 kV main power grid in some areas of South China in 2008 snow disaster
Automation of Electric Power Systems
(2008) - Y. Wang, M. Yang, Y. Zhang, et al., Electromagnetic induction coupling power supply unit of high-voltage transmission...
Cited by (0)
Shuning Wei received M.S. degrees in College of Electrical Engineering, Guangxi University, in 2005. She is currently working toward the Ph.D. degree in the College of Electrical and Information Engineering, Hunan University, Changsha, China. Her research interests include robot control technology, neural network control and machine learning.
Yaonan Wang received the B.S. degree in Computer Engineering from East China Science and Technology University (ECSTU) in 1981 and the M.S. and Ph.D. degrees in Electrical Engineering from Hunan University, Changsha, China, in 1990 and 1994, respectively. From 1994 to 1995, he was a Postdoctoral Research Fellow with the Normal University of Defence Technology. From 1981 to 1994, he worked with the ECSTU. From 1998 to 2000 he was a senior Humboldt Fellow in Germany and a visiting professor at the University of Bremen from 2001 to 2004. He has been a Professor at Hunan University since 1995. His research interests are intelligent control and information processing, robot control, image processing, and industrial process control.
Yi Zuo received the Ph.D. degree in Control Science and Engineering from Hunan University, in 2009. He is a visiting scholar in the Department of Applied Mathematics, University of Waterloo from 2008 to 2009. His scientific interests include neural networks and robotic robust control.