Slip Detection and Classification for Grip Control using Multiple Sensory Modalities on a Biomimetic Tactile Sensor

Achieving human-level performance in dexterous grasping tasks will likely require richer tactile sensing than is currently available [1]. Recently, biomimetic tactile sensors, designed to provide more humanlike capabilities, have been developed. These new sensors provide an opportunity to significantly improve the robustness of robotic manipulation. In order to take full advantage of the information available from such sensors, new estimation techniques have to be developed. Since these sensors provide different sensory modalities, one should also focus on how they can be combined in various manipulation tasks. This paper presents two estimation techniques that use different sensory modalities of biomimetic tactile sensors to detect and classify slip events during grasping. In particular, we present a slip detector, which is able to detect slips more than 30ms before it was detected by an IMU accelerometer. In addition, we demonstrate a slip classifier that is able to classify the type of slip based on different skin distortions with over 80% accuracy before an IMU detects that the object is moving.