【摘要】 随着电力行业蓬勃发展，巡检操作机器人逐步投入变电站应用。当前变电站机器人多以巡视观察为主，在操作类作业过程中，存在智能化程度低、环境适应性差等问题，因此深入探究以人机协同为核心的主从遥操作式机器人显得尤为迫切。文中聚焦于现有力反馈遥操作系统存在的轨迹跟踪抖动现象以及力反馈控制方法不够精细等问题，开展了变电站巡检操作机器人主从控制系统的设计与优化工作。首先，面向变电站检修作业，设计了基于力反馈的主从遥操作机器人及其控制系统。其次，研究了主从控制的末端轨迹跟踪中手部颤震信号的抑制策略，增强了轨迹的跟踪性能。此外，针对力反馈技术，在重力平衡、惯性力简化以及摩擦辨识等维度展开动力学模型的精细化研究，并引入导纳控制方法提升了机器人柔顺特性。最后，搭建试验样机平台，于实际场景中进行功能验证。实验结果表明，主从控制系统的轨迹跟踪精度可达2 mm,低通滤波截止频率2 Hz,滞后时间0.2 s,力反馈误差小于5%,与原有系统相比实现显著提升，更出色地达成了人机协同的预期效果。更多还原

【Abstract】 With the vigorous development of the power industry, inspection and operation robots have been gradually put into application in substations.At present, most substation robots are mainly used for patrol and observation. However, problems such as insufficient intelligence level and poor environmental adaptability exist. Therefore, it is particularly urgent to conduct in-depth research on master-slave teleoperated robots with human-robot collaboration as the core. This paper focuses on the problems existing in the current teleoperation systems, such as the trajectory tracking jitter phenomenon and the lack of refinement in force-feedback control methods. Design and optimization work of the master-slave control system is carried out for substation inspection and operation robots. Firstly, a force-feedback-based master-slave teleoperated robot and its control system are designed. Secondly, strategies to suppress hand tremor signals in master-slave trajectory tracking are studied to enhance dynamic tracking performance. In addition, regarding the force-feedback technology, a refined study of the dynamic model is carried out in dimensions such as gravity compensation, inertial force simplification and friction model identification. The admittance control method is also introduced to improve the compliant characteristics of the robot.Finally, a prototype test platform is built to conduct functional verification in actual scenarios. The experimental results show that the trajectory tracking accuracy of the master-slave control system can reach 2 mm, the cut-off frequency of the low-pass filter is 2 Hz, the lag time is 0.2 s, and the force-feedback error is less than 5%. Compared with the original system, significant improvements have been achieved, and human-robot collaboration can be accomplished more excellently.更多还原