基于最小辨识误差的机器人拖动控制系统设计

Design of Robot Drag Control System Based on Minimum Identification Error

传统的拖动示教算法主要通过动力学模型补偿的方式补偿系统的惯性项和重力项,但由于动力学模型的不准确性导致拖动性能变差。基于关节力传感器,提出仅需通过重力辨识得到的模型实现拖动示教控制的效果。对于重力模型辨识,在LSM的基础上设计误差代价函数,根据代价函数得到误差迭代表达式,使模型估计参数进一步迭代收敛于真实值,实现对协作机器人重力模型的精确辨识。同时,基于重力模型设计拖动控制器,实现柔顺拖动控制。最后,通过MATLAB-Simlink联合仿真,并在搭建的协作机器人平台上验证,结果表明提出的辨识方法具有更高的收敛精度,基于重力模型的拖动控制系统具有更高的稳定性和柔顺性。

The traditional drag-teaching algorithm mainly compensates the inertia and gravity terms of the system by means of dynamic model compensation,but the performance of drag-teaching is poor due to the inaccuracy of the dynamic model.Based on the joint force sensor,this paper proposes that only the model obtained by gravity identification can achieve the effect of drag teaching control.For the gravity model identification,the error cost function is designed based on LSM,and the error iterative expression is obtained according to the cost function,so that the model estimation parameters are further iteratively convergent to the true value,and the accurate identification of the collaborative robot gravity model is realized.At the same time,the drag controller is designed based on the gravity model to realize the smooth drag control.Finally,through MATLAB-Simlink co-simulation and verification on the collaborative robot platform built in our laboratory,the results show that the proposed identification method has higher convergence accuracy,and the drag control system based on gravity model has higher stability and flexibility.