Evaluating Risk Estimation Methods and Path Planning for Safe Human- Robot Cooperation

Integration of machine intelligence into industrial robot cells is demanding and yet not properly solved especially, when it comes to safety aspects between man and machine. The state of the art solution is installing safety fences or vision systems which define static areas of allowed entrance or not allowed entrance. Markerless kinematic estimation for industrial robots is a novelty and state of the art in research. Having knowledge about the kinematic configuration of the human body offers a great possibility to define a function which takes the kinematic configurations of the robot and the human as arguments and find a mapping into a degree of believe with respect to the risk for the human. Depending on the risk estimation, the robot may raise or reduce its speed or if necessary re-plan its path. This paper will discuss how to achieve a two threaded fuzzy logic based system for reason of safety guaranteeing method and controlling path planning. Experimental have been carried out and will be discussed.