Slippage Degree Estimation by Using Vision-Based Tactile Sensor for Dexterous Handling

Abstract In this paper, we present a new slippage degree estimation method, where the vision-based tactile sensor is used to measure multi-dimensional force and moment in contact with an object. We proposed a new method to obtain the stick ratio based on a quantization process of the maximum displacement in the stick region. The stick ratio is useful when we detect macroscopic slippage in advance. We show that the proposed method can be applied to wider situations and obtain more accurate estimation of stick ratio than the previous method. We can use the proposed method for the following situations. First, the applied normal force seriously changes or a certain amount of moment applies to the contact surface. Second, a macroscopic slippage occurs and the sensor surface re-establishs the contact. Third, a complex history of contact makes each local behavior of the surface difficult. The image processing method has been proposed to decrease the data variation of the captured CCD camera image and to cope with the wider situations of contact conditions. The usefulness of the proposed method is confirmed through the experimental results.