神经网络在线自学习跟踪控制及其在伺服系统中的应用

针对传统自适应和自校正控制中存在的问题，提出一种基于神经网络的在线自学习控制方法，既做到了对象模型的在线辨识和控制器的在线设计，又避免了神经网络控制方法通常存在的实时控制的困难，使复杂系统的在线学习控制成为可能。仿真表明该方法具有良好的鲁棒性和控制精度。

Neural networks based on-line self-learning tracking control and its application to servo system

An on - line self - learning control method basedon neural networks, with respect to the problems in the conventionaladaptive and self-learning control, is presented in this paper Thecontrol architecture can modify the model of the model of the controlled plant on-line by introducing a neural network as plant movel It can carryout on - line regulation of the controller independently from the realonntrol operation. Therefore, The real-tine control difficult oftenexisted in neural networks based control methods can be avoidedthus making the on - line learning control of complex systems possible. Simulation results show it has a good robustness and controlprecision.