非完整移动机器人的神经网络滑模自适应轨迹跟踪控制

针对非完整移动机器人的轨迹跟踪控制问题，提出了一种鲁棒项系数自调整的神经网络滑模自适应控制策略。首先由反推法设计运动学控制器；其次，基于滑模控制设计动力学控制器，利用径向基神经网络(RBF)自适应逼近系统非线性不确定性上界，实现鲁棒项系数自调整，克服了传统滑模控制鲁棒项设计需要已知系统不确定性上界的缺陷，实现了速度跟踪。李亚普诺夫稳定性定理保证了闭环系统的稳定性及跟踪误差的渐近收敛。仿真结果进一步验证了所提方案的可行性。

Adaptive Neural Network Sliding Mode Trajectory Tracking Control for Non holonomic Wheeled Mobile Robots

An adaptive neural sliding mode control strategy with the self-tuning of robust item coefficients is proposed for the trajectory tracking of non-holonomic wheeled mobile robots. Firstly, a kinematic controller is designed by means of backstepping technique. Then, the dynamic controller is proposed based on sliding mode control method, in which the upper bound of the uncertainties is adaptively approximated by RBF neural networks and the robust item coefficients are self-tuned. Thus, the disadvantage of the traditional sliding mode controller, which needs to know the boundary of the system uncertainties in advance, is overcome. By using Lyapunov stability theorem, both the stability of closed-loop system and the asymptotical convergence of tracking errors are ensured. Simulation results further validate the effectiveness of the proposed controller.