Collision avoidance for human-robot interaction distinguishing between static and dynamic obstacles

In this contribution, a Model Predictive Control approach for collision avoidance of (mobile) manipulators is extended to handle static and dynamic obstacles in a different way. This is necessary to allow the robot to move close to known static obstacles like tables and at the same time to maximize the distance to dynamic obstacles like humans to ensure safety. However, in both cases collision avoidance has to be guaranteed.