1 Introduction Optimal control problems for complex dynamical systems or dynamic environments can become difficult to solve. In particular, nonconvex pure state constraints, as used in collision avoidance, can make the problem challenging or not real-time capable. Examples can be found in the context of space debris removal [1], where satellites need to execute a collision-free docking maneuver, or in autonomous vehicle racing [2], where one has to overtake other road users.
