Autonomous Collision Avoidance and Separation Assurance for Small UAVs Operating in the National Airspace System

Unmanned air vehicles (UAVs) are presently restricted from operating within the vast National Airspace System (NAS) covering most of North America. The proposed effort seeks to develop FAA-certifiable collision avoidance and separation assurance algorithms that guarantee safe interoperation of small UAVs and other aircraft in the NAS. Small UAVs pose additional size, weight, and power (SWaP) limitations that make this problem especially difficult. The proposed effort adapts and integrates two novel technologies. A rapid collision avoidance algorithm first developed by MIT for unmanned ground vehicles is adapted and unified with an efficient hybrid D* path re-planning algorithm developed previously by the authors. These technologies are chosen specifically for: (1) their robust real-time performance using very limited computation; (2) their accommodation of substantial sensor, disturbance, and modeling uncertainty; and (3) their generation of maneuvers that are guaranteed to be dynamically feasible. The proposed concept is designed to handle dynamic obstacle aircraft that may or may not be cooperative or well-sensed. Because the collision-avoidance and path-planning components both generate complete, realizable trajectories (rather than waypoints), tracking errors can be bounded and aircraft separation can be assured with the highest levels of confidence.