协作机器人位姿误差自适应滑模抑制系统设计

Design of Adaptive Sliding Mode Suppression System for Pose Error of Collaborative Robots

摘要:

协作机器人关节副微观间隙会引发非几何误差累积与时变漂移,使得末端位姿预测置信度降低,其缺乏对多源扰动耦合的有效抑制机制造成补偿响应滞后,使得传统误差模型难以适应真实工况下的精度要求。基于此,设计一种协作机器人位姿误差自适应滑模抑制系统。通过高精度编码器、六维力/力矩传感器与激光跟踪仪构建多源感知层,实现关节微位移与末端实际位姿的同步获取,为误差建模提供可靠数据基础；基于感知数据,采用引入附加旋转参数的改进DH建模方法建立关节间隙-末端误差的显式映射关系,通过非线性微分处理准确描述误差传递机制,解决间隙耦合下的模型失准问题；利用Levenberg-Marquardt优化算法对运动学参数进行高置信度标定,显著提升模型预测一致性；在此基础上,结合PID与自适应滑模复合算法生成实时补偿指令,驱动三向柔性微动机构实现精调与锁紧,以此显著抑制末端位姿误差,实现了协作机器人末端误差的有效补偿。测试结果显示:设计系统应用后预测末端位姿误差与测量末端位姿误差的偏差、末端位姿与目标位姿的误差极小,末端误差补偿指令延迟率整体低于10%,能够满足协作机器人的末端作业精度需求。

Abstract:

The microscopic clearance of the joint pair of collaborative robots can cause the accumulation of non-geometric errors and time-varying drift, reducing the confidence of the end pose prediction. The lack of an effective suppression mechanism for the coupling of multi-source disturbances causes a lag in compensation response, making it difficult for traditional error models to meet the accuracy requirements under real working conditions. Based on this, an adaptive sliding mode suppression system for the pose error of collaborative robots is designed. By using high-precision encoders, six axis force/torque sensors, and laser trackers to construct a multi-source perception layer, synchronous acquisition of joint micro displacement and actual end effector pose can be achieved, providing a reliable data foundation for error modeling; Based on perceptual data, an improved DH modeling method introducing additional rotation parameters is adopted to establish an explicit mapping relationship between joint gap and end effector error. The error transfer mechanism is accurately described through nonlinear differential processing to solve the problem of model misalignment under gap coupling; Using the Levenberg Marquardt optimization algorithm to calibrate kinematic parameters with high confidence significantly improves the consistency of model predictions; On this basis, a real-time compensation instruction is generated by combining PID and adaptive sliding mode composite algorithm to drive the three-way flexible micro motion mechanism for precise adjustment and locking, thereby significantly suppressing the end pose error and achieving effective compensation of the end effector error of the collaborative robot. The test results show that after the application of the designed system, the deviation between the predicted end pose error and the measured end pose error, as well as the error between the end pose and the target pose, are extremely small. The overall delay rate of the end error compensation instruction is less than 10%, which can meet the precision requirements of the collaborative robot"s end operation.