递归型神经网络在机械手逆模学习控制中的应用研究
|
摘要
本文对递归型神经网络在机械手逆模学习控制中的应用进行了研究。传统的基于前馈型神经网络的机械手逆模学习控制中,将机械手的动力学逆模型看作非线性静态映射,利用前馈型神经网络对非线性函数的逼近能力对其进行辨识,所需被测量较多,网络结构复杂,计算量大。如果将机械手的逆模型看作动态系统,并利用递归型神经网络对动力学系统的逼近能力对其进行辨识,则可以减少被测量,简化网络结构。本文使用递归型二阶神经网络对机械手逆动力学模型进行逼近,设计了基于递归型二阶神经网络的机械手逆模学习控制系统,对该系统进行了仿真,并与基于前馈型神经网络的系统进行了比较,以明确递归型神经网络对机械手逆模型的逼近能力,以及将递归型神经网络应用于机械手逆模学习控制的可行性和优越性。
仿真实验表明,本文所设计的递归型二阶神经网络进行离线训练后对机械手逆动力模型有较好的逼近能力,学习效率要优于前馈型神经网络,因此更适用于机械手的在线学习控制;在轨迹控制效果方面,与基于前馈型神经网络的机械手逆模学习控制相比控制精度相差不大;在输出控制量的品质方面,基于递归型二阶神经网络的控制器输出力矩连续稳定,没有强烈的跳动,而基于前馈型神经网络的控制器输出力矩连续性较差,有强烈的突变和跳动;在控制器计算的复杂程度方面,递归型二阶神经网络的结构更简单,所需输入信息更少;在系统模型发生变化时,基于递归型神经网络的控制系统更加稳定。
A type of recurrent neural network named second order recurrent neural networkis applied to the inverse model learning control of manipulators. In the traditionalinverse model learning control method, the model of the manipulator is seen as anonlinear static function and approximated by feed forward neural networks. Thedisadvantages of this method are that much of information of the controlled systemhas to be measured, the size of the networks is large and the computation burden isheavy. In this paper the inverse model of the manipulator is seen as a dynamic systemand approximated by recurrent neural networks, so as to downsize the networks,reduce the requisite information and lessen the computation burden. An inverse modellearning control system of manipulators based on second order recurrent neuralnetworks is proposed and simulated to confirm the feasibility and advantages of theapplication of the recurrent neural networks.
Simulation experiments show that comparing with the method based on feedforward neural network, the advantages of the method presented in this paper are thatthe recurrent neural network is more efficient in offline approximation of the inversemodel of the manipulator, which implies its online control performance will be better;the control precision is not of much difference;the control signal is continuouswithout jerkiness;the size of the network is smaller and the computation burden islighter;when the controlled object model is changing, the control system is morestable.
仿真实验表明,本文所设计的递归型二阶神经网络进行离线训练后对机械手逆动力模型有较好的逼近能力,学习效率要优于前馈型神经网络,因此更适用于机械手的在线学习控制;在轨迹控制效果方面,与基于前馈型神经网络的机械手逆模学习控制相比控制精度相差不大;在输出控制量的品质方面,基于递归型二阶神经网络的控制器输出力矩连续稳定,没有强烈的跳动,而基于前馈型神经网络的控制器输出力矩连续性较差,有强烈的突变和跳动;在控制器计算的复杂程度方面,递归型二阶神经网络的结构更简单,所需输入信息更少;在系统模型发生变化时,基于递归型神经网络的控制系统更加稳定。
A type of recurrent neural network named second order recurrent neural networkis applied to the inverse model learning control of manipulators. In the traditionalinverse model learning control method, the model of the manipulator is seen as anonlinear static function and approximated by feed forward neural networks. Thedisadvantages of this method are that much of information of the controlled systemhas to be measured, the size of the networks is large and the computation burden isheavy. In this paper the inverse model of the manipulator is seen as a dynamic systemand approximated by recurrent neural networks, so as to downsize the networks,reduce the requisite information and lessen the computation burden. An inverse modellearning control system of manipulators based on second order recurrent neuralnetworks is proposed and simulated to confirm the feasibility and advantages of theapplication of the recurrent neural networks.
Simulation experiments show that comparing with the method based on feedforward neural network, the advantages of the method presented in this paper are thatthe recurrent neural network is more efficient in offline approximation of the inversemodel of the manipulator, which implies its online control performance will be better;the control precision is not of much difference;the control signal is continuouswithout jerkiness;the size of the network is smaller and the computation burden islighter;when the controlled object model is changing, the control system is morestable.
引文
[1]Lilai Yan,C. James Li,Robot learning control based on recurrent neural network inverse model,Journal of Robotic Systems 14(3)1997,199–212
[2] Pramod Gupta, Naresh K. Sinha,Intelligent control of robotic manipulators: experimental study using neural networks,Mechatronics,10 (2000),289-305
[3] H. Daniel Pati?o,Ricardo Carelli,Neural networks for advanced control of robot manipulators,Transaction on Neural Networks, Vol. 13, No. 2, March 2002,703-715
[4]M.O. Efe, O. Kaynak,A comparative study of soft-computing methodologies in identification of robotic manipulators,Robotics and Autonomous Systems,30 (2000),221–230
[5]Ziauddin S. M,A. M. S. Zalzala,Model-based compensation and Comparison of Neural Network for uncertainties of robotic arms,IEE Proceedings of Control Theory Applications,Vol.5,1995,501-507
[6]Tso. S. K,Y. H. Fung,Analysis and real-time implementation of a radial-basis function neural-network compensator for high-performance robot manipulators,Mechatronics,Vol.10,2000,265-287
[7]Frank L. Lewis,Kai Liu,Neural Net Robot Controller with Guaranteed Tracking Performance,IEEE Transaction on Neural Networks, Vol. 6, No. 3, MAY 1995,703-715
[8] Frank.L.Lewis,Neural network control of robot manipulators,IEEE Expert,1996,64-75
[9]Rong-Jong Wai,Tracking control based on neural network strategy for robot manipulator,Neurocomputing ,51 (2003),425–445 [10]M.M. Abdelhameed,Adaptive neural network based controller for robots,Mechatronics ,9(1999),147-162
[11]Feng,Gang,A compensation scheme for robot tracking based on networks,Robotics and Autonimous Systems,Vol.15,1995,199-206
[12]Tsay T. I. J,J. H. Huang,Robust nonlinear control of robot manipulator,Proceedings of IEEE International conference on robotics and automation,1994,2083-2094
[13]Glasius,Roy,Neural network dynamics for path planning and obstacle avoidance,Vol.8,No.1,1995,125-133
[14] Wen Chen,Recurrent neural networks applied to robotic motion control,Thesis of Master of Applied Science,University of Toronto,2002
[15]Elias B.Kosmatopoulos, Marios M.Polycarpou, High-order neural network structures for identification of dynamical systems,IEEE Transaction on Neural Nerworks,Vol.6,No.2,March 1995,422-431
[16] Wen Yu,Xiaoou Li.,Some new results on system identification with dynamic neural networks,Transaction on Neural Networks,, Vol. 12, No. 2, March 2001, 412-417
[17] Amir F. Atiya,Alexander G. Parlos,New results on recurrent network training: unifying the algorithms and accelerating Convergence,Transaction on Neural Nerworks, Vol. 11, No. 3, MAY 2000, 697-709
[18]Mbede,Jean Bosco,Robust fuzzy and reccurent neural network motion control among dynamic obstacles for robot manipulators,Proceedings of IEEE International Conference on Robotics and Automation,Vol.3,2000,2136-2145
[19]Nerandra,Kumpati S,Identification and control of dynamical systems usiong neural networks,IEEE Transaction on Neural Networks, Vol. 1, No. 1, March 1990,4-27
[20]George A. Rovithakis,Tracking control of multi-input affine nonlinear dynamical systems with unknown nonlinearities using dynamical neural networks,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 29, No. 2, April 1999,179-189 动
[21] George A. Rovithakis,M. A. Christodoulou,Adaptive control of unknown plants using dynamical neural networks,Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 24, No. 3, April 1994,400-412
[22] George A. Rovithakis,M. A. Christodoulou,Adaptive control with recurrent high-order neural network: theory and industrial applications,London: Springer,2000
[23]Shen,Jie,A class of modified hopfield networks for control of linear and nonlinear systems,Proceedings of American Control Conference,Vol.2,1998,964-969
[24]王耀南,机器人智能控制工程,北京:科学出版社,2004,193-203
[25]王灏,毛宗源,机器人的智能控制方法,北京:国防工业出版社,2002,1-3
[26]张铁,谢存禧,机器人学,广州:华南理工出版社,2002,93-111
[27]徐丽娜,神经网络控制,哈尔滨:哈尔滨工业大学出版社,2003
[28]蔡自兴,机器人学,北京:清华大学出版社,2000,89-126
[29]Simon Haykin,Neural Netwoks:A Comprehensive Foundation,Prentice-Hall,Inc,1999,33-40,483-623
[30] A.G. Parlos,O.T. Rais, A.F. Atiya,Multi-step-ahead prediction using dynamic recurrent neural networks,Neural Networks 13 (2000) 765-786
[31]Hao Jianbin,Shaohua,One step ahead predictive control of nonlinear systems by neural networks,Proceedings of IEEE International joint conference on neural networks,1993,2761-2764
[32]Jin Liang,P. N. Nikiforuk,Approximation of discrete-time state-space trajectories using dynamic recurrent neural networks,IEEE Transactions on Automatic Control,Vol.40,No.7,July 1995,1266-1270
[33] Jun Wang, Qingni Hu,A lagrangian network for kinematic control of redundant robot manipulators, Transaction on Neural Nerworks, Vol. 10, No. 5, March 1999,1123-1132
[34] Wai Sum Tang,Jun Wang,Two recurrent neural networks for local joint torque optimization of kinematically redundant manipulators,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 30, No. 1, February 2000,120-128
[35]Jin,Liang,Universal approximation using dynamic recurrent neural networks:discrete-time version,Proceedings of IEEE International Conference on Neural Network,Vol.1,1995,403-408
[36] Wai Sum Tang,Jun Wang,A recurrent neural network for minimum infinity-norm kinematic control of redundant manipulators with an improved problem formulation and reduced architecture complexity,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 31, No. 1, February 2001,98-105
[37]伍朝晖,谢光汉,符曦,巧解机器人动力学方程—MATLAB语言应用一例,电脑与信息技术,1997年第5期,23-25
[38]千敏耀,马骏骑,基于MATLAB的PUMA机器人运动学仿真,昆明理工大学学报(理工版),第28卷第6期,2003年12月,50-53
[39]杨艳丽,史维祥,机械手分散CMAC控制器,机械科学与技术,第20卷第3期,2001年5月,536-538
[40]孙袆,王耀南,模糊小波基神经网络的机器人轨迹跟踪控制,控制理论与应用,第20卷第1期,2003年2月,49-53
[41]杨国军,崔平远,基于神经网络的机械手鲁棒输出跟踪控制,哈尔滨工业大学学报,第34卷第5期,2002年10月,628-672
[42]关新平,唐英干,段广仁,机械手臂基于神经网络动态补偿的自适应控制,控制理论与应用,第20卷第1期,2003年2月,102-105
[43] Ya-Chen Hsu,Guanrong Chen,A fuzzy adaptive variable structure controller with applications to robot manipulators , IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 31, No. 3,June 2001,331-340
[44] Yang Gao,Meng Joo Er,,Adaptive control of robot manipulators using fuzzy neural networks, IEEE Transaction on Electronics,Vol. 48, No. 6, December 2001,1274-1278
[45]谭震,蒋洪明,采用神经元补偿器的机械手计算转矩控制算法,机器人,第21卷第二期,1999年3月,104-109
[46]李士勇, 模糊控制、神经控制和智能控制论,哈尔滨:哈尔滨工业大学出版社 1998,3-20
[47]何玉彬,神经网络控制技术及其应用,北京:科学出版社 2000
[48]Kim,Sung-Woo,and Ju-Jang Lee,Filter-error-learning neural network for stable trajectory tracking control of robot manipulators ,Mechatronics,Vol.6,No.6,1996,181-192
[49]Poanyal A.S.,Wen Yu,Nonlinear adaptive trajectory tracking using dynamic neural networks,IEEE Transaction on Neural Networks, Vol. 10, No. 6, November 1999,1402-1411
[50] 薛定宇,陈阳泉,基于MATLAB/simulink的系统仿真技术与应用,北京:清华大学出版社,2002
[51] [法]莫克塔里(Mokhtari,m.)等著,赵彦玲等译 ,MATLAB与SIMULINK工程应用,北京:科学出版社,2002
[53] 施阳 严卫生 李俊 等,MATLAB语言精要及动态仿真工具SIMULINK,西安:西安电子可以大学出版社,1997
[54]魏克新等,MATLAB语言与自动控制系统设计,北京:机械工业出版社,1997
[55] 陈永春编著,从Matlab/Simulink模型到代码实现,西安:西安电子科技大学出版社,2002
[56] 张志涌等,精通 MATLAB 6.5版,西安:西安电子科技大学出版社,2003
[2] Pramod Gupta, Naresh K. Sinha,Intelligent control of robotic manipulators: experimental study using neural networks,Mechatronics,10 (2000),289-305
[3] H. Daniel Pati?o,Ricardo Carelli,Neural networks for advanced control of robot manipulators,Transaction on Neural Networks, Vol. 13, No. 2, March 2002,703-715
[4]M.O. Efe, O. Kaynak,A comparative study of soft-computing methodologies in identification of robotic manipulators,Robotics and Autonomous Systems,30 (2000),221–230
[5]Ziauddin S. M,A. M. S. Zalzala,Model-based compensation and Comparison of Neural Network for uncertainties of robotic arms,IEE Proceedings of Control Theory Applications,Vol.5,1995,501-507
[6]Tso. S. K,Y. H. Fung,Analysis and real-time implementation of a radial-basis function neural-network compensator for high-performance robot manipulators,Mechatronics,Vol.10,2000,265-287
[7]Frank L. Lewis,Kai Liu,Neural Net Robot Controller with Guaranteed Tracking Performance,IEEE Transaction on Neural Networks, Vol. 6, No. 3, MAY 1995,703-715
[8] Frank.L.Lewis,Neural network control of robot manipulators,IEEE Expert,1996,64-75
[9]Rong-Jong Wai,Tracking control based on neural network strategy for robot manipulator,Neurocomputing ,51 (2003),425–445 [10]M.M. Abdelhameed,Adaptive neural network based controller for robots,Mechatronics ,9(1999),147-162
[11]Feng,Gang,A compensation scheme for robot tracking based on networks,Robotics and Autonimous Systems,Vol.15,1995,199-206
[12]Tsay T. I. J,J. H. Huang,Robust nonlinear control of robot manipulator,Proceedings of IEEE International conference on robotics and automation,1994,2083-2094
[13]Glasius,Roy,Neural network dynamics for path planning and obstacle avoidance,Vol.8,No.1,1995,125-133
[14] Wen Chen,Recurrent neural networks applied to robotic motion control,Thesis of Master of Applied Science,University of Toronto,2002
[15]Elias B.Kosmatopoulos, Marios M.Polycarpou, High-order neural network structures for identification of dynamical systems,IEEE Transaction on Neural Nerworks,Vol.6,No.2,March 1995,422-431
[16] Wen Yu,Xiaoou Li.,Some new results on system identification with dynamic neural networks,Transaction on Neural Networks,, Vol. 12, No. 2, March 2001, 412-417
[17] Amir F. Atiya,Alexander G. Parlos,New results on recurrent network training: unifying the algorithms and accelerating Convergence,Transaction on Neural Nerworks, Vol. 11, No. 3, MAY 2000, 697-709
[18]Mbede,Jean Bosco,Robust fuzzy and reccurent neural network motion control among dynamic obstacles for robot manipulators,Proceedings of IEEE International Conference on Robotics and Automation,Vol.3,2000,2136-2145
[19]Nerandra,Kumpati S,Identification and control of dynamical systems usiong neural networks,IEEE Transaction on Neural Networks, Vol. 1, No. 1, March 1990,4-27
[20]George A. Rovithakis,Tracking control of multi-input affine nonlinear dynamical systems with unknown nonlinearities using dynamical neural networks,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 29, No. 2, April 1999,179-189 动
[21] George A. Rovithakis,M. A. Christodoulou,Adaptive control of unknown plants using dynamical neural networks,Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 24, No. 3, April 1994,400-412
[22] George A. Rovithakis,M. A. Christodoulou,Adaptive control with recurrent high-order neural network: theory and industrial applications,London: Springer,2000
[23]Shen,Jie,A class of modified hopfield networks for control of linear and nonlinear systems,Proceedings of American Control Conference,Vol.2,1998,964-969
[24]王耀南,机器人智能控制工程,北京:科学出版社,2004,193-203
[25]王灏,毛宗源,机器人的智能控制方法,北京:国防工业出版社,2002,1-3
[26]张铁,谢存禧,机器人学,广州:华南理工出版社,2002,93-111
[27]徐丽娜,神经网络控制,哈尔滨:哈尔滨工业大学出版社,2003
[28]蔡自兴,机器人学,北京:清华大学出版社,2000,89-126
[29]Simon Haykin,Neural Netwoks:A Comprehensive Foundation,Prentice-Hall,Inc,1999,33-40,483-623
[30] A.G. Parlos,O.T. Rais, A.F. Atiya,Multi-step-ahead prediction using dynamic recurrent neural networks,Neural Networks 13 (2000) 765-786
[31]Hao Jianbin,Shaohua,One step ahead predictive control of nonlinear systems by neural networks,Proceedings of IEEE International joint conference on neural networks,1993,2761-2764
[32]Jin Liang,P. N. Nikiforuk,Approximation of discrete-time state-space trajectories using dynamic recurrent neural networks,IEEE Transactions on Automatic Control,Vol.40,No.7,July 1995,1266-1270
[33] Jun Wang, Qingni Hu,A lagrangian network for kinematic control of redundant robot manipulators, Transaction on Neural Nerworks, Vol. 10, No. 5, March 1999,1123-1132
[34] Wai Sum Tang,Jun Wang,Two recurrent neural networks for local joint torque optimization of kinematically redundant manipulators,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 30, No. 1, February 2000,120-128
[35]Jin,Liang,Universal approximation using dynamic recurrent neural networks:discrete-time version,Proceedings of IEEE International Conference on Neural Network,Vol.1,1995,403-408
[36] Wai Sum Tang,Jun Wang,A recurrent neural network for minimum infinity-norm kinematic control of redundant manipulators with an improved problem formulation and reduced architecture complexity,IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 31, No. 1, February 2001,98-105
[37]伍朝晖,谢光汉,符曦,巧解机器人动力学方程—MATLAB语言应用一例,电脑与信息技术,1997年第5期,23-25
[38]千敏耀,马骏骑,基于MATLAB的PUMA机器人运动学仿真,昆明理工大学学报(理工版),第28卷第6期,2003年12月,50-53
[39]杨艳丽,史维祥,机械手分散CMAC控制器,机械科学与技术,第20卷第3期,2001年5月,536-538
[40]孙袆,王耀南,模糊小波基神经网络的机器人轨迹跟踪控制,控制理论与应用,第20卷第1期,2003年2月,49-53
[41]杨国军,崔平远,基于神经网络的机械手鲁棒输出跟踪控制,哈尔滨工业大学学报,第34卷第5期,2002年10月,628-672
[42]关新平,唐英干,段广仁,机械手臂基于神经网络动态补偿的自适应控制,控制理论与应用,第20卷第1期,2003年2月,102-105
[43] Ya-Chen Hsu,Guanrong Chen,A fuzzy adaptive variable structure controller with applications to robot manipulators , IEEE Transaction on Systems, Man, and Cybernetis—Part B: Cybernetis, Vol. 31, No. 3,June 2001,331-340
[44] Yang Gao,Meng Joo Er,,Adaptive control of robot manipulators using fuzzy neural networks, IEEE Transaction on Electronics,Vol. 48, No. 6, December 2001,1274-1278
[45]谭震,蒋洪明,采用神经元补偿器的机械手计算转矩控制算法,机器人,第21卷第二期,1999年3月,104-109
[46]李士勇, 模糊控制、神经控制和智能控制论,哈尔滨:哈尔滨工业大学出版社 1998,3-20
[47]何玉彬,神经网络控制技术及其应用,北京:科学出版社 2000
[48]Kim,Sung-Woo,and Ju-Jang Lee,Filter-error-learning neural network for stable trajectory tracking control of robot manipulators ,Mechatronics,Vol.6,No.6,1996,181-192
[49]Poanyal A.S.,Wen Yu,Nonlinear adaptive trajectory tracking using dynamic neural networks,IEEE Transaction on Neural Networks, Vol. 10, No. 6, November 1999,1402-1411
[50] 薛定宇,陈阳泉,基于MATLAB/simulink的系统仿真技术与应用,北京:清华大学出版社,2002
[51] [法]莫克塔里(Mokhtari,m.)等著,赵彦玲等译 ,MATLAB与SIMULINK工程应用,北京:科学出版社,2002
[53] 施阳 严卫生 李俊 等,MATLAB语言精要及动态仿真工具SIMULINK,西安:西安电子可以大学出版社,1997
[54]魏克新等,MATLAB语言与自动控制系统设计,北京:机械工业出版社,1997
[55] 陈永春编著,从Matlab/Simulink模型到代码实现,西安:西安电子科技大学出版社,2002
[56] 张志涌等,精通 MATLAB 6.5版,西安:西安电子科技大学出版社,2003