高精度陀螺稳定跟踪系统神经网络预测控制

摘要

针对高精度陀螺稳定跟踪系统，提出了一种基于小波基函数神经网络的非线性系统的一步超前预测控制算法。该方法利用小波网络学习非线性系统，利用小波神经网络模型作为系统的预测模型，控制信号直接通过极小化期望输出值与预测输出值之间的偏差来获得。通过对陀螺稳定跟踪控制系统的仿真，表明该算法具有优良的控制品质。%In this paper, a one-step ahead predictive control algorithm based on wavelet neural networks is proposed for the gyro stability tracking system with high accuracy. The wavelet neural networks is used to learn the unknown nonlinear system. Taking the wavelet neural networks as a predictive model of the system, the control signals are directly obtained by minimizing the deviation between the expected value and the output value of the predictive model. The simulation result shows that the proposed method can give good control performance.