一类自适应神经网络控制器的设计

Design of adaptive neural network controller for a class of multiple time-delays nonlinear system

针对一类具有下三角形式的多重时滞非线性系统,研究了基于小波神经网络的自适应控制器设计方案。应用后推(backstepp ing)设计方法在递推的每一步设计一个Lyapunov-Krasovsk ii泛函并在最后一步得出控制u。应用Lyapunov-Krasovskii泛函稳定性定理证明了该控制器能保证闭环系统是半全局最终一致有界的(SGUUB)。仿真实例验证了该控制器的有效性。

An adaptive wavelet neural network （WNN） controller was designed for a class of parametric strict-feedback multipie time-delays nonlinear system. Based on the backstepping technique, a Lyapunov-Krasovskii functional is constructed at each step and the controller u is obtained at the last step. By using the law of Lyapunov-Krasovskii functional stability, the semi-global uniform ultimate boundedness （SGUUB） of all the signals in the closed-loop system was proved. At last, the feasibility of this controller was illustrated by a simulation example.