Including intent in detect-and-avoid systems for remotely piloted aircraft systems

Abstract

Current standards for detect-and-avoid (DAA) systems of remotely piloted aircraft systems (RPAS) use state data like (relative) position and speed for the provision of guidance to remain well clear (RWC) with air traffic. They do not use intent data for planned route towards a destination. This paper proposes an intent-based DAA system that uses an A* path planning approach. The performance of this A* DAA system is compared with the ACAS Xu standard for a set of horizontal encounters between RPAS pairs. This comparison is done for deterministic and stochastic encounter-scenarios, which account for sensor errors and variation in remote pilot performance. The results show that in the ACAS Xu scenarios the aircraft can attain livelock conditions that prevent them from reaching their destination. Also, in these encounter-scenarios often loss of DAA well clear (LDWC) conditions occur, and they are sensitive for sensor errors and closed-loop delays. The simulation results of the A* DAA encounter-scenarios are without livelock or LDWC conditions and the variations due to sensor errors and closed-loop delays are limited. They show that the intent-based A* DAA system is a promising approach for more effective DAA.