Initial iterative learning control strategy based on CMAC neural networks

Abstract

In order to avoid the problem of slow converging speed and long time expenditure when traditional iterative learning control system faced a new environment an improved algorithm was proposed to obtain the initial value of the iterative learning control based on CMAC neural networks.Desired control input of iterative learning control that was estimated by CMAC neural networks and feedback PID networks based on the historical control experience worked as the initial control input of the iterative learning control.With the role of open-closed loop P-type iterative learning control algorithm the actual output trajectory of the system could track desired trajectory in accurate requirements using less iteration.The simulation results of the robotic system show the algorithm is feasible and effective.