论文信息 - Neural networks for modelling and control of a non-linear dynamic system

Neural networks for modelling and control of a non-linear dynamic system

The authors describe the use of neural nets to model and control a nonlinear second-order electromechanical model of a drive system with varying time constants and saturation effects. A model predictive control structure is used. This is compared with a proportional-integral (PI) controller with regard to performance and robustness against disturbances. Two feedforward network types, the multilayer perceptron and radial-basis-function nets, are used to model the system. The problems involved in the transfer of connectionist theory to practice are discussed.<<ETX>>

[1] D. Sbarbaro,et al. A nonlinear receding horizon controller based on connectionist models , 1991, [1991] Proceedings of the 30th IEEE Conference on Decision and Control.

[2] F. Girosi,et al. Networks for approximation and learning , 1990, Proc. IEEE.

[3] I. J. Leontaritis,et al. Input-output parametric models for non-linear systems Part II: stochastic non-linear systems , 1985 .

[4] John Moody,et al. Fast Learning in Networks of Locally-Tuned Processing Units , 1989, Neural Computation.

[5] James S. Albus,et al. New Approach to Manipulator Control: The Cerebellar Model Articulation Controller (CMAC)1 , 1975 .

[6] James S. Albus,et al. Data Storage in the Cerebellar Model Articulation Controller (CMAC) , 1975 .

[7] Hecht-Nielsen. Theory of the backpropagation neural network , 1989 .

[8] Brian Armstrong. On finding 'exciting' trajectories for identification experiments involving systems with non-linear dynamics , 1987, Proceedings. 1987 IEEE International Conference on Robotics and Automation.