Target-Oriented Fuzzy-Collision-Avoidance for Vehicles

Autonomous mobile systems (AMS) mostly navigate on paths created by global path planners. But due to often appearing unforeseen objects the AMS needs a local collision avoidance (LCA) which takes over control functions and guides the AMS around obstacles. Mostly this maneuver results in leaving the preplanned path. Unfortunately it is possible that the AMS cannot find back to the precomputed path because the obstacles' arrangement is too complex. Therefore it is useful to equip the LCA with a target-oriented component during collision avoidance in order to guide the AMS back to the precomputed path. Our approach realizes a fuzzy-based LCA using direct sensor information to generate a steering angle. The results of our examinations show, that the AMS successfully reaches the target-point in most of the cases; without any global path planning, only with the means of fuzzy logic and a target-oriented component.