Slipping control through tactile sensing feedback

The paper presents a novel slipping control algorithm based on the exploitation of the tactile sensor integrated into the DEXMART anthropomorphic robotic hand. The Extended Kalman Filter (EKF) is used to solve in real-time the nonlinear model of the sensor, allowing to estimate the contact geometry variables and the friction coefficient. The innovative proposed slipping control is based on the use of the estimation error of the EKF as an indicator of incipient slipping events. The control algorithm computes the suitable grip force based on the estimation error. The effectiveness of the proposed approach is shown with experimental results.