A Pose Estimation and Optimization Method for Mobile Manipulator’s End-effectors Based on Stereo Vision and ICoP Algorithm

The accurate pose estimation of the mobile manipulator’s end-effector is pivotal in mobile machining of large and complex components. In this paper, a pose tracking method based on binocular stereo vision in large-scale workspace is presented. The binocular camera and numbers of planar coded markers are used to estimate the pose of the mobile manipulator’s end-effector. Specifically, the binocular camera measures its relative pose by simultaneously photographing the fiducial markers on the ground and the tracking markers on the end-effector. Through the calibration of the fiducial markers’ absolute position, the pose of the end-effector is obtained. Besides, considering the influence of visual measurement error and calibration error on pose estimation accuracy, a robust high-accuracy poses estimation method based on iterative corresponding point optimization (ICoP) algorithm is proposed to deal with the influence of measurement error on pose estimation accuracy, which is more robust to data with large errors. Simulated comparison with the existing Chan’s algorithm and corresponding point algorithm is also studied to verify the superiority of the proposed method. A practical experiment was also performed and the results that the poses estimation accuracy for the mobile manipulator’s end-effector are ±0.215 mm and ±0.016° within the range of 6 m × 4 m × 1.5 m, respectively, demonstrates the proposed methods are effective.