Hybrid A* path search with resource constraints and dynamic obstacles

Alán Cortez, Bryce Ford, Indranil Nayak, Sriram Narayanan, Mrinal Kumar

Abstract: In recent years, there has been considerable interest in incorporating complex constraints into path-planning for autonomous vehicles so as to expand their scope of operation in challenging environments. This paper considers two forms of complex constraints for path planning of a Dubins agent: 1) time-varying obstacles that are capable of translation, rotation, as well as changing their shape, and 2) resource constraints that exert a path-dependent cumulative load (or damage) on the vehicle that must be maintained under a stipulated threshold. In the context of unmanned aerial vehicles (UAVs), examples of the former include weather phenomena, such as storms, turbulence and ice. In the case of resource constraints, the agent is allowed to “fly through” the obstacle, so long as the total accumulated damage incurred along its path is kept under a stipulated threshold. Examples of such path-dependent loading constraints include flight of a UAV over a wildfire, wherein the resource constraint represents placing an upper limit on the increase in temperature along the agent’s path to keep onboard.