CPG仿生控制研究
|
摘要
仿生控制的研究对于机器人技术的进展有着重大的意义。近年来,模拟生物中枢模式发生器(CPG)的仿生控制已经成为仿生控制领域的研究热点之一。CPG能够在缺乏高层控制信号和外部反馈的情况下,自发产生稳定的节律性运动,这一运动具有易被高层命令调节、耦合性强、适应性强、结构简单等优点,可以减少控制系统的工作量,节约工作时间。目前,这一仿生控制的研究在国际上尚处于初级阶段,特别在结构的简单化、高层调节机制的加入等方面还不能很好地模拟生物的特性,充分发挥这一控制的优点。
本文根据基于振动理论的三神经元CPG局部神经元网络数学模型建立了一种新的CPG控制模型,并针对机器人双关节的控制,以CPG控制模型为主体建立了双关节仿生控制系统。
首先,根据三神经元CPG局部神经元网络数学模型,模拟生物神经系统中各神经元的响应原理,建立了各神经元的控制模型,分别实现节律运动的发生、反馈信息的传导与被控对象的驱动三项功能,并在仿真平台上对模型的控制特性做出验证。仿真结果表明,控制模型可以较好地模拟生物神经元的特性。
其次,引入了仿生控制系统的概念,模拟生物的神经-肌肉控制系统建立了CPG仿生控制系统。通过CPG规划器、驱动神经元前项通道、反馈神经元通道的建立实现整个系统的运动模式发生、控制对象驱动与外界信息反馈,并通过这三个部分的连接与相互作用实现运动的简单高层控制信息与反馈信息调节。
然后,模拟生物神经的整合作用,在单关节运动控制系统的基础上建立了机器人双关节耦合控制系统。通过控制系统自抑制系数的改变,实现了单CPG对双关节的耦合控制,并可通过简单高层控制信息对此运动进行调节。
最后,应用此系统进行了单关节控制、双关节耦合控制与反射作用模拟等仿真实验。实验结果表明,此系统可较好地模拟生物特性,完成双关节耦合节律性运动的发生,并能通过较为简单的结构实现运动的高层调节以及生物反射作用的模拟。
Bionic control is of great significance for the progress of robot technology. In recent years, bionic control based on central pattern generator (CPG) has become one of the hot spots of the field of bionic control. CPG can produce movement of spontaneous steady rhythm which is susceptible to high-level orders regulating, has strong coupling, multi-mode, strong adaptability and simple structure to the lack of high-level external control signal and the feedback, it can reduce the workload of the control system, saving hours of work. Currently, the research of the bionic control of CPG in the international arena is still at the preliminary stage. The control systems can not well simulate biological characteristics nor give full play to the advantage of this control method, particularly in the aspect of simplified structure, movement of Multi-mode and to follow high-level mechanism.
Based on the three local neurons network model of CPG adopted the theory of vibration, we established a new control system model of CPG. Aimed at the control of a single leg of two joints of the robot, we built a single legged bionic control system of two joints which contains CPG as a primary part.
First, We built the models of each neuron in the control system simulating the response theory of neurons in the biological nervous system according to the three local neurons network model of CPG adopted the theory of vibration. The neuron models achieve the rhythmic motion, feedback to the conduction and the object-driven respectively. Then validated the control characteristics on the simulation platform. Simulation results show that the model can simulate the biological characteristics of neurons preferably.
Secondly, the paper simulated biological nerve-muscle control system to establish a single joint multi-loop bionic control system. We established the producing of the movement mode, the driving of the objects, the feeding back of the information by building the planner, the forward path and the feedback path, and established the interaction of high-level campaign to achieve control and feedback regulation with this three-part connection.
Then, with the simulation of biological neural integration, this paper built a coupling control system of a single robot leg of two joints based on the single joint control system. The system can realize coupling control of two joints using only one CPG, with the adoption of high-level multi-regulating control system.
Finally, we conducted a single joint control, a coupling control of two joints and simulation of reflection with the application of this system. Experimental results show that this system can accomplish double coupling joints rhythm movements simulating biological characteristics, and can achieve movements changing, high-level regulation of the movements and reflection simulation with a relatively simple structure.
本文根据基于振动理论的三神经元CPG局部神经元网络数学模型建立了一种新的CPG控制模型,并针对机器人双关节的控制,以CPG控制模型为主体建立了双关节仿生控制系统。
首先,根据三神经元CPG局部神经元网络数学模型,模拟生物神经系统中各神经元的响应原理,建立了各神经元的控制模型,分别实现节律运动的发生、反馈信息的传导与被控对象的驱动三项功能,并在仿真平台上对模型的控制特性做出验证。仿真结果表明,控制模型可以较好地模拟生物神经元的特性。
其次,引入了仿生控制系统的概念,模拟生物的神经-肌肉控制系统建立了CPG仿生控制系统。通过CPG规划器、驱动神经元前项通道、反馈神经元通道的建立实现整个系统的运动模式发生、控制对象驱动与外界信息反馈,并通过这三个部分的连接与相互作用实现运动的简单高层控制信息与反馈信息调节。
然后,模拟生物神经的整合作用,在单关节运动控制系统的基础上建立了机器人双关节耦合控制系统。通过控制系统自抑制系数的改变,实现了单CPG对双关节的耦合控制,并可通过简单高层控制信息对此运动进行调节。
最后,应用此系统进行了单关节控制、双关节耦合控制与反射作用模拟等仿真实验。实验结果表明,此系统可较好地模拟生物特性,完成双关节耦合节律性运动的发生,并能通过较为简单的结构实现运动的高层调节以及生物反射作用的模拟。
Bionic control is of great significance for the progress of robot technology. In recent years, bionic control based on central pattern generator (CPG) has become one of the hot spots of the field of bionic control. CPG can produce movement of spontaneous steady rhythm which is susceptible to high-level orders regulating, has strong coupling, multi-mode, strong adaptability and simple structure to the lack of high-level external control signal and the feedback, it can reduce the workload of the control system, saving hours of work. Currently, the research of the bionic control of CPG in the international arena is still at the preliminary stage. The control systems can not well simulate biological characteristics nor give full play to the advantage of this control method, particularly in the aspect of simplified structure, movement of Multi-mode and to follow high-level mechanism.
Based on the three local neurons network model of CPG adopted the theory of vibration, we established a new control system model of CPG. Aimed at the control of a single leg of two joints of the robot, we built a single legged bionic control system of two joints which contains CPG as a primary part.
First, We built the models of each neuron in the control system simulating the response theory of neurons in the biological nervous system according to the three local neurons network model of CPG adopted the theory of vibration. The neuron models achieve the rhythmic motion, feedback to the conduction and the object-driven respectively. Then validated the control characteristics on the simulation platform. Simulation results show that the model can simulate the biological characteristics of neurons preferably.
Secondly, the paper simulated biological nerve-muscle control system to establish a single joint multi-loop bionic control system. We established the producing of the movement mode, the driving of the objects, the feeding back of the information by building the planner, the forward path and the feedback path, and established the interaction of high-level campaign to achieve control and feedback regulation with this three-part connection.
Then, with the simulation of biological neural integration, this paper built a coupling control system of a single robot leg of two joints based on the single joint control system. The system can realize coupling control of two joints using only one CPG, with the adoption of high-level multi-regulating control system.
Finally, we conducted a single joint control, a coupling control of two joints and simulation of reflection with the application of this system. Experimental results show that this system can accomplish double coupling joints rhythm movements simulating biological characteristics, and can achieve movements changing, high-level regulation of the movements and reflection simulation with a relatively simple structure.
引文
1田文罡.多足步行机器人控制系统的研究. 2005年3月,华中科技大学硕士论文
2张秀丽,郑浩峻,段广洪.动物运动控制机理及对机器人控制的启示.机器人技术与应用. 2002(5): 24~26
3张秀丽,郑浩峻,陈恳,段广洪.机器人仿生学研究综述.机器人. 2002(2): 188~192
4喻宗泉.神经控制的应用与展望.自动化与仪器仪表. 2000(4): 1~2
5顾毅.智能控制发展综述.信息技术. 2000(6): 39~40
6蔚东晓,贾霞彦.模糊控制的现状与发展.自动化与仪器仪表. 2006(6): 4~7
7顾毅.智能控制发展综述.信息技术. 2000(6): 39~40
8刘宇.模糊控制的现状与发展.油气田地面工程. 2006(9): 13~21
9刘曙光,王志宏,费佩燕,王斌.模糊控制的发展与展望.机电工程. 2000(1): 9~12
10靖稳峰,魏红,段惠娣.遗传算法及其发展现状.西安工业学院学报. 2000(3): 230~235
11张丽萍,柴跃廷.遗传算法的现状及发展动向.信息与控制. 2001(6): 531~536
12蔡自兴,周翔,李枚毅,雷鸣.基于功能/行为集成的自主式移动机器人进化控制体系结构.机器人. 2000年(3): 169~175
13 Brewer B G, Jung R.“Sensitivity analys is of a hybrid neural network for locomotor control in the lamprey”[J]. IEEE.Biomedical Engineering. 1997: 353- 356
14 Selverston AI,Rowat PF.Modeling phase synchronization in a biologically-based network.In:Neural Net-works.IJCNN,1992.396
15 Bj?rn Brembs. Operant conditioning in invertebrates. Current Opinion in Neurobiology. Volume 13. Issue 6. 2003, 9.: 710-717
16 Delcomyn F. Neural basis of rhythmic behavior in animals. [J]. Science. 1980, 210: 492~498
17 Matsugu M. Entrainment, Instability, Quasi-periodicity, and Chaos in a Compound Neural Oscillator" [J]. Journal of Computational Neuroscience. 1998,5: 35-51
18 Sandro Nicole, Eliano Pessa. A network model with auto-oscillating output and dynamic connections. [J]. Biological Cybernetics. 1994, 70(3): 275~280
19 BeerRD, ChielHJ, GallagherJC. Journal of Computational Neuro science, 1999, 7:119
20 Matsugu M.Journal of Computational Neuro science, 1998, 5:35
21 Beer RD,Chiel HJ,Gallagher JC.Journal of Computational Neuro science, 1999,7:119
22 Hooper SL.CurrentBiology,2000,10(5):R17
23 K. Matsuoka. Mechanisms of frequency and pattern control in the neural rhythm generators. Biological Cybernetics, 56: 345-353, 1987
24 William Bialek, Fred Rieke, Rob R. De. Ruyter van Steveninck, David Warland, Reading a Neural Code, Science, Jun 28, 1991, 252, 5014, 1854-1857
25 HILLEL J. CHIEL, RANDALL D. BEER, JOHN C. GALLAGHER. Evolution and Analysis of Model CPGs forWalking: I. Dynamical Modules. Journal of Computational Neuroscience 7, (1999): 99–118
26 Franca F.M.G, Yang Zhijun, IEEE Neural Networks, 2000, 4: 290
27 A simple legged locomotion gaint model,S.T.Venkataraman,Robotics and Autonomous Systems 22 (1997): 75-85
28 Miyakoshi S,Taga G,Kuniyoshi Y,etal./"Three dimensional bipedal stepping motion using neural oscillators-towards humanoid motionin the real world "[A]. IEEE/RSJ.Intelligent Robots and Systems [C]. 1998.84-89
29 Kimura,H. Fukuoka,Y. Adaptive dynamic walking of a quadruped robot on irregular terrain by using neural system model. Intelligent Robots and Systems. 2000.10.31~11.5(2): 979-984
30 Hiroshi Kimura, Yasuhiro Fukuoka, Hiroyuki Nakamura. Biologically Inspired Adaptive Dynamic Walking of the Quadruped on Irregular Terrain. Intelligent Robots and Systems 3 (1999): 1657~1662
31 Fukuoka, Y. , Kimura, H. , Hada, Y. , Takase, K. Adaptive dynamic walking of a quadruped robot 'Tekken' on irregular terrain using a neural system model. Robotics and Automation. 2003, 9(2): 2037~2042
32 Kimura, H. Fukuoka, Y. Biologically inspired adaptive dynamic walking in outdoor environment using a self-contained quadruped robot: 'Tekken2'.Intelligent Robots and Systems. 2004, 9(1): 986 - 991
33 Shinkichi Inagaki. Hideo Yuasa. Tamio Arai. CPG model for autonomous decentralized multi-legged robot system—generation and transition of oscillation patterns and dynamics of oscillators. Robotics and Autonomous Systems 44 (2003) 171–179
34 Or. J. , Takanishi. A. A biologically inspired CPG-ZMP control system for the real-time balance of a single-legged belly dancing robot. Intelligent Robots and Systems. 2004, 9, 28: 931~936
35 Tenore. F. , Etienne-Cummings. R. , Lewis, M.A. A programmable array of silicon neurons for the control of legged locomotion. Circuits and Systems. 2004, 5, 23~26(5). V-349~V-352
36 Tenore. F. , Vogelstein. R.J. , Etienne-Cummings. R. , Cauwenberghs. G. , Lewis. M.A. , Hasler. P. A spiking silicon central pattern generator with floating gate synapses [robot control applications]. Circuits and Systems. 2005, 5, 23~26(4): 4106~4109
37 Arena. P. , Fortuna. L. , Frasca. M. , Patane. L. , Pavone. M. Climbing obstacles via bio-inspired CNN-CPG and adaptive attitude control. Circuits and Systems. 2005,5,23~26(5): 5214~5217
38 Zhang,Z.G. Fukuoka,Y. Kimura,H. Adaptive Running of a Quadruped Robot Using Delayed Feedback Control. Robotics and Automation. 25.4.18~22 3739~3744
39 Lewis,M.A. Tenore,F. Etienne-Cummings,R. CPG Design using Inhibitory Networks. Robotics and Automation. 2005,4,18~22: 3682~3687
40 Gen Endo, Jun Nakanishi, Jun Morimoto, Gordon Cheng. Experimental Studies of a Neural Oscillator for Biped Locomotion with QRIO. Robotics and Automation. 2005.4.18~22. 596~602
41郑浩峻,张秀丽,李铁民,段广洪.基于CPG原理的机器人运动控制方法. 高技术通讯. 2003(7): 64~68
42郑浩峻,张秀丽,关旭,汪劲松.基于生物中枢模式发生器原理的四足机器人.清华大学学报(自然科学版)2004,44(2): 166~169
43 Long Wang, Shuo Wang, Zhiqiang Cao, Min Tan, Chao Zhou, Haiquan Sang, Zhizhong Shen. Motion Control of a Robot Fish Based on CPG. Industrial Technology. 2005,12,14~17. 1263~ 1268
44 Zhenli Lu,Shugen Ma,Bin Li,Yuechao Wang. Serpentine locomotion of a snake-like robot controlled by cyclic inhibitory CPG model. Intelligent Robots and Systems. 2005, 8, 2~6. 96~101
45寿天德.神经生物学.高等教育出版社. 2001. 6
2张秀丽,郑浩峻,段广洪.动物运动控制机理及对机器人控制的启示.机器人技术与应用. 2002(5): 24~26
3张秀丽,郑浩峻,陈恳,段广洪.机器人仿生学研究综述.机器人. 2002(2): 188~192
4喻宗泉.神经控制的应用与展望.自动化与仪器仪表. 2000(4): 1~2
5顾毅.智能控制发展综述.信息技术. 2000(6): 39~40
6蔚东晓,贾霞彦.模糊控制的现状与发展.自动化与仪器仪表. 2006(6): 4~7
7顾毅.智能控制发展综述.信息技术. 2000(6): 39~40
8刘宇.模糊控制的现状与发展.油气田地面工程. 2006(9): 13~21
9刘曙光,王志宏,费佩燕,王斌.模糊控制的发展与展望.机电工程. 2000(1): 9~12
10靖稳峰,魏红,段惠娣.遗传算法及其发展现状.西安工业学院学报. 2000(3): 230~235
11张丽萍,柴跃廷.遗传算法的现状及发展动向.信息与控制. 2001(6): 531~536
12蔡自兴,周翔,李枚毅,雷鸣.基于功能/行为集成的自主式移动机器人进化控制体系结构.机器人. 2000年(3): 169~175
13 Brewer B G, Jung R.“Sensitivity analys is of a hybrid neural network for locomotor control in the lamprey”[J]. IEEE.Biomedical Engineering. 1997: 353- 356
14 Selverston AI,Rowat PF.Modeling phase synchronization in a biologically-based network.In:Neural Net-works.IJCNN,1992.396
15 Bj?rn Brembs. Operant conditioning in invertebrates. Current Opinion in Neurobiology. Volume 13. Issue 6. 2003, 9.: 710-717
16 Delcomyn F. Neural basis of rhythmic behavior in animals. [J]. Science. 1980, 210: 492~498
17 Matsugu M. Entrainment, Instability, Quasi-periodicity, and Chaos in a Compound Neural Oscillator" [J]. Journal of Computational Neuroscience. 1998,5: 35-51
18 Sandro Nicole, Eliano Pessa. A network model with auto-oscillating output and dynamic connections. [J]. Biological Cybernetics. 1994, 70(3): 275~280
19 BeerRD, ChielHJ, GallagherJC. Journal of Computational Neuro science, 1999, 7:119
20 Matsugu M.Journal of Computational Neuro science, 1998, 5:35
21 Beer RD,Chiel HJ,Gallagher JC.Journal of Computational Neuro science, 1999,7:119
22 Hooper SL.CurrentBiology,2000,10(5):R17
23 K. Matsuoka. Mechanisms of frequency and pattern control in the neural rhythm generators. Biological Cybernetics, 56: 345-353, 1987
24 William Bialek, Fred Rieke, Rob R. De. Ruyter van Steveninck, David Warland, Reading a Neural Code, Science, Jun 28, 1991, 252, 5014, 1854-1857
25 HILLEL J. CHIEL, RANDALL D. BEER, JOHN C. GALLAGHER. Evolution and Analysis of Model CPGs forWalking: I. Dynamical Modules. Journal of Computational Neuroscience 7, (1999): 99–118
26 Franca F.M.G, Yang Zhijun, IEEE Neural Networks, 2000, 4: 290
27 A simple legged locomotion gaint model,S.T.Venkataraman,Robotics and Autonomous Systems 22 (1997): 75-85
28 Miyakoshi S,Taga G,Kuniyoshi Y,etal./"Three dimensional bipedal stepping motion using neural oscillators-towards humanoid motionin the real world "[A]. IEEE/RSJ.Intelligent Robots and Systems [C]. 1998.84-89
29 Kimura,H. Fukuoka,Y. Adaptive dynamic walking of a quadruped robot on irregular terrain by using neural system model. Intelligent Robots and Systems. 2000.10.31~11.5(2): 979-984
30 Hiroshi Kimura, Yasuhiro Fukuoka, Hiroyuki Nakamura. Biologically Inspired Adaptive Dynamic Walking of the Quadruped on Irregular Terrain. Intelligent Robots and Systems 3 (1999): 1657~1662
31 Fukuoka, Y. , Kimura, H. , Hada, Y. , Takase, K. Adaptive dynamic walking of a quadruped robot 'Tekken' on irregular terrain using a neural system model. Robotics and Automation. 2003, 9(2): 2037~2042
32 Kimura, H. Fukuoka, Y. Biologically inspired adaptive dynamic walking in outdoor environment using a self-contained quadruped robot: 'Tekken2'.Intelligent Robots and Systems. 2004, 9(1): 986 - 991
33 Shinkichi Inagaki. Hideo Yuasa. Tamio Arai. CPG model for autonomous decentralized multi-legged robot system—generation and transition of oscillation patterns and dynamics of oscillators. Robotics and Autonomous Systems 44 (2003) 171–179
34 Or. J. , Takanishi. A. A biologically inspired CPG-ZMP control system for the real-time balance of a single-legged belly dancing robot. Intelligent Robots and Systems. 2004, 9, 28: 931~936
35 Tenore. F. , Etienne-Cummings. R. , Lewis, M.A. A programmable array of silicon neurons for the control of legged locomotion. Circuits and Systems. 2004, 5, 23~26(5). V-349~V-352
36 Tenore. F. , Vogelstein. R.J. , Etienne-Cummings. R. , Cauwenberghs. G. , Lewis. M.A. , Hasler. P. A spiking silicon central pattern generator with floating gate synapses [robot control applications]. Circuits and Systems. 2005, 5, 23~26(4): 4106~4109
37 Arena. P. , Fortuna. L. , Frasca. M. , Patane. L. , Pavone. M. Climbing obstacles via bio-inspired CNN-CPG and adaptive attitude control. Circuits and Systems. 2005,5,23~26(5): 5214~5217
38 Zhang,Z.G. Fukuoka,Y. Kimura,H. Adaptive Running of a Quadruped Robot Using Delayed Feedback Control. Robotics and Automation. 25.4.18~22 3739~3744
39 Lewis,M.A. Tenore,F. Etienne-Cummings,R. CPG Design using Inhibitory Networks. Robotics and Automation. 2005,4,18~22: 3682~3687
40 Gen Endo, Jun Nakanishi, Jun Morimoto, Gordon Cheng. Experimental Studies of a Neural Oscillator for Biped Locomotion with QRIO. Robotics and Automation. 2005.4.18~22. 596~602
41郑浩峻,张秀丽,李铁民,段广洪.基于CPG原理的机器人运动控制方法. 高技术通讯. 2003(7): 64~68
42郑浩峻,张秀丽,关旭,汪劲松.基于生物中枢模式发生器原理的四足机器人.清华大学学报(自然科学版)2004,44(2): 166~169
43 Long Wang, Shuo Wang, Zhiqiang Cao, Min Tan, Chao Zhou, Haiquan Sang, Zhizhong Shen. Motion Control of a Robot Fish Based on CPG. Industrial Technology. 2005,12,14~17. 1263~ 1268
44 Zhenli Lu,Shugen Ma,Bin Li,Yuechao Wang. Serpentine locomotion of a snake-like robot controlled by cyclic inhibitory CPG model. Intelligent Robots and Systems. 2005, 8, 2~6. 96~101
45寿天德.神经生物学.高等教育出版社. 2001. 6