Hierarchical Control System For Flexible Materials Handling Robots Using Neural Network

This paper presents hierarchical control system for compensation & vibration control of flexible material handling robot using neural network for t he control parameters scheduling and dynamic control. In this paper, compensation control method is classified into two cases; (i) a method based on estimated value of tip position, and (ii) a method based on measured value of tip position. The method (i) is only available for the simple cases (i.e., the deflection is small and the section of the material is unity). In the method (ii), by choosing control parameters properly, it is shown that the compensation & vibration control with small position error is possible. However, it is quite difficult to select proper control parameters for each different payloads so that the control system remains stable. To solve this problem, we propose hierarchical control system. This system has a gain scheduler at the upper level to roughly select proper control parameters for each different materials, and has a neural network controller at the lower level to improve control performance. After designing gain scheduler and neural network controller properly, the system can adapt to improve the response time for the large number of different kinds of materials.

[1]  Kazuhiro Kosuge,et al.  New strategy for hierarchical intelligent control of robotic manipulator-hybrid neuromorphic and symbolic control , 1991, [Proceedings] 1991 IEEE International Joint Conference on Neural Networks.

[2]  H. Hablani.  Constrained and unconstrained modes - Some modeling aspects of flexible spacecraft , 1982 .

[3]  D. Graham,et al.  The synthesis of "optimum" transient response: Criteria and standard forms , 1953, Transactions of the American Institute of Electrical Engineers, Part II: Applications and Industry.

[4]  Kazuhiro Kosuge,et al.  Neuromorphic control for robotic manipulator: hierarchical hybrid neuromorphic control system , 1992 .

[5]  Fumihito Arai,et al.  Hierarchical hybrid neuromorphic control system for robotic manipulator , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[6]  Fumihito Arai,et al.  Neuromorphic sensing and control-impact control of robotic manipulator , 1991, [Proceedings] 1991 IEEE International Joint Conference on Neural Networks.

[7]  Michael I. Jordan.  Supervised learning and systems with excess degrees of freedom , 1988 .

[8]  正東 馬,et al.  Trajectory Control and Inverse Dynamics of Flexible Robot Arms , 1988 .

[9]  Bruno Siciliano,et al.  Trajectory control of a non-linear one-link flexible arm , 1989 .

[10]  Toshio Fukuda,et al.  Neural Network Applications for Robotic Motion Control , 1990, J. Robotics Mechatronics.

[11]  Zheng-Dong Ma,et al.  Inverse dynamics of flexible robots , 1989 .