Tactile Sensors Based Object Recognition and 6D Pose Estimation

For robots working in real world environments, especially in the underwater area, it is necessary to achieve robust recognition and 6D pose estimation of freely standing movable objects using tactile sensors. Until now this problem remains unsolved due to the limited capability of the available tactile sensors. However, in our research group we have developed a highly reliable tactile sensor system, with high spatial and force resolution [2]. Such a sensor system enables us to achieve robust recognition and 6D localization of static objects as well as freely standing movable objects in high noise conditions which can be expected underwater. In this paper we will present our approaches and simulation based results.