摘要
使用改进的神经网络模型结构与算法来辨识未知非线性系统,具有辨识精度高,速度快的特点。该方法简单有效,为设计非线性对象控制器提供了一条思路,从而摆脱了用线性模冲近似被控对象的粗略做法。算法中,学习率采用随误差变化率而改变的做法减小了学习率选取的盲目性．加速了网络训练过程。
A modified neural network model and an algorithm which were used to identify an unkown nonlinear system were presented. This is of fast and accurate feature. It was so simple that supplied an idea for designing nonlinear object control device. The learning rate is variable according to the error change of identificationFurthen more it sped up the learning procedure.
出处
《河北理工学院学报》
2006年第4期64-67,共4页
Journal of Hebei Institute of Technology
