基于模型不确定补偿的轮式移动机器人反演复合控制

针对轮式移动机器人存在模型不确定性、非线性以及未建模的动态特性等因素,严重影响系统轨迹跟踪的稳定性和精确性,提出一种基于系统模型不确定性补偿的反演复合控制策略。基于非完整轮式移动机器人的运动学模型,采用反演控制思想以及李雅普诺夫稳定性判据设计轨迹跟踪的虚拟速度控制量,作为系统的持续激励输入。考虑轮式移动机器人具有模型不确定性和外部有界力矩干扰,根据轮式移动机器人的动力学模型推导得到系统不确定项,并采用具有高度非线性拟合特性的神经网络对其估计,得到模型的力矩控制量,且由李雅普诺夫稳定性分析得到不确定项的自适应律,实现自调整和实时轨迹跟踪。对比仿真表明,该复合控制策略能自适应的跟踪期望轨迹,与单一的反演控制、模型不确定性补偿控制策略、传统PID控制相比,均具有更好的鲁棒性和高的跟踪精度。

The composite control of backstepping control based on uncertain model compensation of wheeled mobile robot

Given these factors of model uncertainty, non-linearity and unmodeled dynamic characteristics existing in wheeled mobile robots, which seriously affected the stability and accuracy of trajectory tracking, a backstepping composite control strategy based on model uncertainty compensation was proposed. Based on the kinematics model of a nonholonomic wheeled mobile robot, backstepping control and Lyapunov stability criterion were adopted to design virtual velocity control quantity as continuous incentive input for trajectory tracking. Considering the model uncertainty and external bounded moment disturbance of wheeled mobile robots, the uncertainties of the system were derived from the dynamic model of wheeled mobile robots, and the moment control quantity of model was acquired by using the neural network with highly nonlinear fitting characteristics, and then adaptive law of uncertainties was obtained from Lyapunov stability analysis to realize self-adjustment and real-time trajectory tracking. The simulation results showed that the proposed composite control strategy could track the reference trajectory adaptively, and had better robustness and tracking accuracy than the single backstepping control strategy, model uncertainty compensation control strategy and PID controller.