一种无模型的机器人非线性输出反馈控制

A model-free nonlinear output feedback controller for robot manipulators

针对非线性不确定机器人系统的轨迹跟踪控制问题,在传统PID控制器的基础上,设计带观测器的无模型连续控制器.控制器由PD控制部分和非线性积分部分组成,其中PD控制部分用于镇定系统,非线性积分部分用于提高算法的收敛速度,同时考虑到速度信号不易获取,引入线性观测器估计机器人关节的速度,并在理论上证明了观测误差和跟踪误差最终趋向于零.在SCARA机器人的两个旋转关节上进行轨迹跟踪试验,结果表明,采用该控制算法得到了理想的控制性能,在同样的反馈增益下,采用观测器后的平均跟踪误差和最大跟踪误差降低了30%以上,而控制输出的震荡程度只有无观测器时的60%.

A continuous model-free controller with a linear observer based on the classical PID controller was proposed for trajectory tracking of robot manipulators with nonlinear uncertainties.The controller is composed of a PD and nonlinear integral controllers.The PD control term was used to stabilize the closed-loop system trajectories while the nonlinear integrator to improve the convergence.By considering the difficulty in getting the velocity data,the linear observer was proposed to estimate the actual joint velocity.It was proven that the tracking errors and observation errors tend to zero theorytically.The experimental results on the two rotating joints of SCARA robot manipulator show that the presented controller performed well and the average and maximum tracking errors with observer were reduced by more than 30%.The degree of control chattering is only 60% of that without observer under the same feedback gain.