Design of nonlinear servo controller based on the performance function and approximate solution by neural network

摘要

This paper is concerned with a performance based nonlinear servo controller. The aim of usual nonlinear servomechanism problem was to asymptotically stabilize the closed loop system, i.e., to let the error converge to zero. Here we try to improve the convergence speed by designing a servo controller so as to decrease a performance function like the squared error momently. At the same time, we propose a new method applying neural network to solve approximately the nonlinear output regulation equation which is necessary to be solved as designing the nonlinear servo controller. The effectiveness of the proposed method was confirmed with simulation results for TORA model.