A novel accurate positioning method for object pose estimation in robotic manipulation based on vision and tactile sensors

The high positioning accuracy of robotic systems is fundamental and critical for dexterous and precise manipulation. In this paper, a novel accurate positioning method based on vision and tactile sensors for object pose estimation in robotic manipulation is proposed to improve the positioning accuracy of robotic systems. The proposed methodology mainly includes two stages of vision-based rough positioning stage and tactile-based precise positioning stage which makes full use of the multimodal information for positioning and object pose estimation in robotic systems. In the vision-based rough positioning stage, the position and orientation of the target object can be obtained by RGB and depth images, and the robot approaches the target object considering a safe distance. In the tactile-based precise positioning stage, the tactile images are captured by the designed tactile sensor through the touch movement of the robot, and the position and orientation of the target object can be precisely calculated further to complete the positioning process. Finally, the experiments have been conducted to verify the proposed method in a real application platform, and achieved positioning accuracies of 0.138 mm and 0.203° have shown great improvement of 96.1% and 93.9% from 3.552 mm and 3.343° which can meet the demands for robotic precise positioning and manipulation.