Hybrid Vision-force Control for Human Robot Co-carrying on Ball and Board Systems

We propose a hybrid vision-force control that combines vision servo control with force control to perform a robot collaborative by human carrying task on a ball-and-board systems. The proposed control allows the robot to effectively cooperate with human by successfully moving the board to the desired target position, while actively avoiding the ball from falling off the board. A RBFNN-based force control is designed to ensure the robot is able of complying with human motion under uncertain robot dynamics. Visual servo control based on image recognition and location is used to generate reference height and angle to adjust the position and posture of the board to keep the ball on the board. The experiment illustrates that the proposed technique is practical and effective to perform human-robot co-carrying on a ball-and-board systems.