论文部分内容阅读
本文提出一种动态线性或非线性系统的神经网络逆模型辨识结构．并引出两种PID与神经网络逆模型相结合的自适应控制方案．神经网络模型采用基于U－D分解卡尔曼滤波学习算法（UDK）的动态前向多层同．仿真结果表明了所述辨识方案的有效性及特点．
In this paper, a neural network inverse model identification structure of dynamic linear or nonlinear system is proposed. And leads to two PID and neural network inverse model of the combination of adaptive control scheme. The neural network model uses dynamic forward multilayers based on the U-D Decomposition Kalman Filter Learning Algorithm (UDK). The simulation results show the effectiveness and characteristics of the identification scheme.