Dynamic obstacle avoidance
Clark R. Karr, Michael A. Craft, Jaime E. Cisneros · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 1995
Techniques to allow simulated entities to avoid static terrain, such as trees, buildings, rivers, etc., have been in use in Distributed Interactive Simulation (DlS) environments for many years. Avoidance of objects in motion, “dynamic obstacles”, is much more complex. Although simple dynamic obstacle collision avoidance has been implemented in other systems, the resulting behavior is usually less than realistic. The Institute for Simulation and Training (1ST) has investigated techniques to allow simulated entities to make reasonably intelligent and realistic maneuvers intended to avoid dynamic (and static) objects. The goal has been to find new methods which will yield improved collision avoidance behavior without excessive computational cost. As a result of this work, 1ST has developed a novel approach to attack the DIS dynamic obstacle avoidance (DOA) problem by combining two disparate motion planning approaches: potential fields and grid based route planning.