全向移动机器人的神经滑模自适应PD跟踪控制

Neural Sliding Mode Adaptive PD Tracking Control of Omnidirectional Mobile Robots

针对四轮全向机器人数学模型中未知因数的干扰,提出了一种基于神经滑模的比例微分(Proportional plus Derivative,PD)自适应跟踪控制策略。首先,对机器人的动力学模型进行分析;其次,利用反步轨迹跟踪控制器得到虚拟速度,基于虚拟输入与实际输入的误差设计PD控制器;接着,采用神经网络来逼近动力学模型中的不确定性,并利用Lyapunov定理证明了系统的稳定性。仿真和实验结果表明,该控制策略具有有效性。

A proportional plus derivative(PD)adaptive tracking control strategy based on neural sliding mode is proposed for the interference of unknown factors in the mathematical model of four-wheel omnidirectional robots.Firstly,the dynamics model of the robot is analyzed.Secondly,the backstepping trajectory tracking controller is used to obtain the virtual velocity,and the PD controller is designed based on the error between the virtual input and the actual input.Then,the neural network is used to approximate the uncertainty in the dynamics model,and the stability of the system is proved by the Lyapunov theorem.Simulation and experimental results show that the control strategy is effective.