带有死区输入和未建模动态的神经网络控制

With a dead zone input unmodeled dynamics and neural network control

针对一类具有死区非线性输入和未建模动态的非线性系统,提出一种自适应神经网络控制方法。该方法将后推技术和动态面技术结合,克服了计算复杂性问题,放宽了动态不确定性的假设,取消了神经网络逼近误差有界。借助中值定理和Young’s不等式,保证整个设计只需一个自适应参数,且控制增益只需存在一个上、下界。理论分析证明闭环系统所有信号半全局一致终结有界。仿真结果验证所提方案的有效性。

The adaptive neural network controller is proposed for a class of nonlinear systems with dead-zone nonlinear input and unmodeled dynamics. Combining dynamic surface control with backstepping technique, this scheme overcomes the ＂explosion of complexity＂ problem, broadens the variables of unmodeled dynamics and cancels the assumption of the neural network approximation error to be bounded. Using the mean value theorem and Young＇ s inequality, only one adaptive parameter is needed for the whole design, furthermore only one upper and lower bound of control gain are needed to exist. By theoretical analysis, the closed-loop systems is shown to be semi-globally uniformly ultimately bounded, Simulation results demonstrate the effectiveness of the proposed method.