Vision-based space manipulator online self-calibration

Because of great shock in the process of launching of rocket, structure deformation during work, displacement of sensors and other factors, the kinematic parameters of space manipulator will deviate from the parameters calibrated on the ground, which will decrease the positioning accuracy of space manipulator. Therefore, the space manipulator should be calibrated to ensure its positioning accuracy. However, the existing calibration methods are offline and depend on the working environment excessively, so a novel online self-calibration method is presented for kinematic parameters of space manipulator in this paper. This method utilizes the information of position and orientation of a fixed target on space station and adopts rank-one quasi-Newton method to calculate the errors of kinematic parameters, and the position and orientation of the fixed target can be measured by the camera mounted on space manipulator end-effector. This method can calibrate the manipulator parameters online, has low demand for working environment, and can implement parameters calibration for space manipulator without external measurement equipment in space. Simulation results verified the feasibility and effectiveness of the proposed self-calibration method.