一个基于神经网络模型的鲁棒自适应控制算法

对于具有不确定因素的离散非线性动态系统,通过校正神经网络预报器的输出,运用加权 预报控制性能指标和网络辨识器模型局部线性化的思想,提出了一个间接鲁棒自适应神经网 络控制算法,仿真研究证实了该控制策略的鲁棒性和有效性.

A Robust Adaptive Control Algorithm Via Neural Networks Model

In this paper a robust adaptive neural network control algorithm for unknown nonlinear dynamical systems with stable unmodeled dynamics or/and bounded disturbances is proposed. The key ideas are that (1) linearizing the neural network identifier model to obtain an explicit adaptive control law, and (2) correcting the neural network predictive output of the real plant. An example is studied and the simulation results are given to demonstrate the validity of our presented adaptive control strategy.