基于优化ADRC的单臂机器人轨迹跟踪研究

针对机器人在不确定环境下受到内外界挠动的影响，末端执行器不能跟踪已规划的轨迹运行，以Kinova MICO2机器人为例，研究轨迹自抗挠跟踪模型。建立了机器人关节空间运动模型，根据挠动特点构造了二阶ADRC框架模型，解决了末端执行器偏离规划轨迹的问题。分析了机器人ADRC参数对轨迹跟踪的意义，为了确保系统跟踪的动态稳定性，建立GA-RBF网络优化了原始ADRC。最后分析轨迹跟踪前后的离散点绝对误差，验证机器人轨迹跟踪的可靠性。实验结果表明，GA-RBF优化的ADRC使得机器人实际轨迹能较好收敛于理想的规划轨迹，为工业应用中对轨迹要求较高的任务提供算法参考。

Study on Trajectory Tracking of Single-arm Robot Based on Optimized ADRC

Aiming at the fact that robot is affected by the internal and external disturbance in the uncertain environment, and the end effector can‘t track the planned trajectory, Kinova MICO 2 robot is used as an example to study the trajectory active disturbance rejection tracking model. Firstly, the spatial motion model of the robot joint is established, and the second-order ADRC frame model is constructed according to the characteristics of the disturbance, which solves the problem that the end effector deviates from the planned trajectory. Secondly, the significance of ADRC parameters to trajectory tracking is analyzed. In order to ensure the dynamic stability of system tracking, GA_RBF network is established to optimize the original ADRC. Finally, the absolute error of discrete points before and after trajectory tracking is analyzed, and the reliability of robot trajectory tracking is verified by analyzing the error. The experimental results show that the ADRC optimized by GA_RBF makes the actual trajectory of the robot converge to the ideal planning trajectory, which provides algorithm reference for the tasks with high trajectory requirements in industrial applications.