UAV Collision Avoidance Utilizing ADS-B and TCAS with Independent Ranging

Abstract

With the proliferation of Unmanned Aerial Vehicles (UAVs), their integration into the National Airspace System (NAS) is becoming increasingly imperative. To accomplish such integration, UAVs must be able to ''sense and avoid'' other aerial vehicles. Automatic Dependent Surveillance-Broadcast (ADS-B) and Traffic Collision Avoidance Systems (TCAS) offer attractive data for use in estimating the relative position of an intruder aircraft. ADS-B will soon be mandatory in the NAS and its protocol permits independent range measurements in addition to the position and velocity estimates contained in messages. TCAS, while not mandatory for all aircraft, uses active interrogation of aircraft transponders - which are mandatory in a large portion of controlled airspace - to provide accurate range estimates to intruders. ADS-B reports, ADS-B independent ranges, and TCAS ranges were used as measurements to Extended Kalman Filters (EKFs) which provided estimates of the relative position and velocity of an intruder aircraft. Three EKF versions were examined: one with no delay states, one with position delay states to improve relative velocity estimates, and one with velocity delay states to estimate relative aircraft acceleration. The EKF with no delay states provided the most promising results to simulated encounters. When the process noise values were properly tuned, it provided improvements over un-processed ADS-B position and velocity values and performed well during periods of complete signal loss and random message loss.