不确定轮式移动机器人统一自适应神经网络H_∞控制

Unified Adaptive Neural Network H_∞ Control of Uncertain Wheeled Mobile Robots

提出了一种基于自适应神经网络控制和H_∞控制的轮式移动机器人光滑全局跟踪和镇定统一的控制器。首先采用横截函数方法,扩展系统控制输入,建立与原系统等价的、输入输出完全解耦的无奇异全驱动系统,再对新系统设计自适应神经网络H_∞控制器。自适应神经网络控制可有效补偿系统的复杂不确定项。H_∞控制器可同时对系统扰动和神经网络逼近误差进行预定水平抑制,进一步提高控制器的适应性,优化系统的控制性能。仿真结果验证了算法的有效性。

A smooth global unified controller of trajectory tracking and stabilization was proposed for nonholomomic wheeled mobile robots based on adaptive neural network control and H∞ control. Firstly, the system control inputs were expanded by transverse function method, a nonsingular full drive system which was equivalent to original system was established with decoupled input-output. Then an adaptive neural network H∞ controller was designed for the new system, such that the com- plex system uncertainty was compensated effectively by the adaptive neural network. Disturbances and approximation errors were attenuated with a prescribed disturbance lever by the H∞ control. Adapta- bility of the controller were further improved, and the control performance was optimized. The effec- tiveness of the algorithm were verified by simulation results.