基于视觉引导的腕臂机器人柔顺装配方法分析OA

To address challenges such as part reflection,inaccurate positioning,and frequent jamming in the automated assembly of high-speed rail overhead contact line wrist-arms,this paper proposes a robot compliant assembly method integrating vision and force sensing.Through adaptive exposure and deep learning,the method achieves robust part recognition and coarse positioning.Based on 3D point cloud visual servo,it accomplishes sub-millimeter precision positioning and online compensation.During assembly,an impedance model and finite-state machine enable force/position hybrid control,allowing the robot to adaptively perform searching,alignment,and insertion based on real-time torque.Experimental results demonstrate that the proposed method can achieve an assembly success rate of over 92%,significantly outperforming traditional approaches,and provides a reliable technical solution for precision assembly of complex components.