Geometric model-based joint angle selection criterion for force parameter identification & Decoupling control method of position and posture in shaft-hole assembly

Junhe Wang,Yong Jiang,Song Lin,Fanxu Kong

DOI: https://doi.org/10.1109/cyber53097.2021.9588304

2021-07-27

Abstract:Robotic vision servo technology has been developed to control a robotic arm to a specified position with the guidance of a vision sensor. However, for tasks that require high precision control, guidance by vision alone is not sufficient. We need to use compliance control method based on contact force information to accomplish fine-grained operations. In the paper, firstly, a geometric model-based joint angle selection criterion for force parameter identification is proposed, which can improve the stability and accuracy of force sensor parameter identification. Secondly, due to the unstable and too small values of the collected torque information, and in order to omit the tedious force analysis, as well as the costly trial-and-error process, we divide the shaft-hole assembly problem into two parts. One is to combine the point cloud normal vector estimation and posture transformation to achieve the control of the end point position of the axis body constant, while adjusting the posture of the shaft body by speed control; and the other is to use the force feedback damping control to complete the assembly tasks when the shaft body direction is consistent with the hole axis direction. After these two parts of operation, the decoupling of posture and position can be realized, shaft hole assembly at any posture of the workpiece (hole) can be completed. The effectiveness and feasibility of the algorithm is verified by conducting experiments on the UR5 robot.