This paper presents a method for robot self-recognition and self-adaptationthrough the analysis of the contact between the robot end effector and itssurrounding environment. Often, in off-line robot programming, the idealizedrobotic environment (the virtual one) does not reflect accurately the real one.In this situation, we are in the presence of a partially unknown environment(PUE). Thus, robotic systems must have some degree of autonomy to overcome thissituation, especially when contact exists. The proposed force/motion controlsystem has an external control loop based on forces and torques exerted on therobot end effector and an internal control loop based on robot motion. Theexternal control loop is tested with an optimal proportional integrative (PI)and a fuzzy-PI controller. The system performance is validated with real-worldexperiments involving contact in PUEs.
展开▼
