A Novel Detect-and-Avoid Approach for UAS in Urban Environments

This paper addresses one of the key contemporary issues in avionics and Air Traffic Management (ATM) systems research: the development of a certifiable Detect-and-Avoid (DAA) system for trusted autonomous Unmanned Aircraft System (UAS) operations. DAA systems for unmanned aircraft are required to achieve a level of safety that is at least equivalent to the see-and-avoid capabilities of their manned counterparts. To achieve this overarching goal, DAA systems must account for the uncertainty arising from a number of different sources within the available Separation Assurance and Collision Avoidance functionalities. This paper proposes a novel approach to UAS DAA which accounts for the performance of the Communication, Navigation and Surveillance (CNS) systems, as well as a number of dynamic factors impacting the likelihood of a collision. The methodology is underpinned by rigorous modelling of the CNS performance characteristics, and the translation of those characteristics to the spatial domain to form a dynamic (i.e., performance-driven) geo-fence around each aircraft or ground obstacle track. Inflations are then applied to the volume to account for factors such as relative platform dynamics, uncertainty in wind forecasts, wake turbulence and adverse weather conditions. The methodology is demonstrated through case studies involving the use of Primary Surveillance Radar (PSR) and ADS-B to detect intruder aircraft.