论文部分内容阅读
研究了一类基于两层动态神经网的仿射型鲁棒自适应跟踪问题．对于未知的仿射非线性系统，提出了新的鲁棒学习算法，该算法不需要知道理想权值的界，用δ－保护解决了文［２］提出的δ－保护而引起的不连续的问题，从理论上证明了闭环系统的鲁棒稳定性．仿真结果验证了提出的动态网自适应控制算法的有效性
A class of affine robust adaptive tracking problem based on two-layer dynamic neural network is studied. For unknown affine nonlinear systems, a new robust learning algorithm is proposed, which does not need to know the bounds of ideal weights and solves the problem of δ-protection caused by [2] using δ-protection The problem proves the robust stability of the closed-loop system theoretically. The simulation results verify the effectiveness of the proposed dynamic network adaptive control algorithm