论文部分内容阅读
对未知仿射系统提出了用动态神经网实现鲁棒直接自适应控制的策略。基于Lya-punov理论,获得一个稳定并且连续的学习律,闭环系统被证明是鲁棒稳定的。此方法不需要离线学习阶段也不要求初始的参数误差足够小.
A new strategy of robust direct adaptive control based on dynamic neural network is proposed for unknown affine system. Based on the Lya-punov theory, a stable and continuous learning law is obtained, and the closed-loop system proves to be robust and stable. This method does not require an offline learning phase and does not require the initial parameter error to be small enough.