Combining eye-in-hand visual servoing and force control in robotic tasks using the task frame

This paper shows how to use Mason's compliance frame or task frame to model, implement and execute robotic tasks which combine force control and visual servoing in an uncalibrated workspace. Both vision and force sensors are mounted on the manipulator. Hence, the vision algorithms benefit from the known depth to the contact point, when operating under force controlled contact. The sensor data is fused in a hybrid position/force control scheme with internal velocity controller. The fusing method depends on the high level task description. Experimental results are presented to validate the approach