带有非线性输入和动态不确定性的神经网络控制

摘要

A scheme for design of adaptive neural network controller is proposed for the class of non-affine systems with dead-zone nonlinear input and unmodeled dynamics. Compared with the existing methods, this scheme can improve the processing effect for unmodeled dynamics, simplify the design procedure and increase the tracking precision. Furthermore, it can make the stability analysis more reasonable. 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.%针对一类带有非线性输入和未建模动态的非仿射系统，提出一种自适应控制器的设计方案。与现有结果相比，此方案改进了对未建模动态的处理方法，简化了设计过程，提高了跟踪精度，使得稳定性分析更加合理。李亚普诺夫方法分析证明该方案能够保证所有信号半全局一致终结有界。仿真结果验证了所提方案的有效性。