论文部分内容阅读
本文基于提出的递归多层神经网络结构,进行了非线性系统的模型参考自适应控制研究,并为其提供了全解耦的推广Kalman滤波器学习算法,通过SISO非线性动力学系统的控制仿真,表明本文的控制方法对于阶未知的非线性对象或具有未知延迟的系统控制是可行的。
Based on the proposed recursive multi-layer neural network structure, this paper studies the model reference adaptive control of nonlinear systems, and provides a fully decoupled extended Kalman filter learning algorithm. Through the control simulation of SISO nonlinear dynamics system , Which indicates that the control method in this paper is feasible for the control of order unknown nonlinear objects or system with unknown delay.