Detect and Avoid (DAA) Alerting Performance Comparison: CPDS vs. ACAS-Xu

摘要

Detect and Avoid (DAA) systems are integral to an Unmanned Aircraft System's (UAS) ability to Remain Well Clear (RWC) of other aircraft; they are an enabling technology for UAS to integrate into the National Airspace System (NAS). “Minimum Operational Performance Standards (MOPS) for Detect and Avoid (DAA) Systems”¹(RTCA DO-365) specifies criteria for determining when an alerting algorithm must, may, and must not issue an alert in an encounter with an intruding aircraft. Various organizations have developed prototype alerting algorithms for DAA, including National Aeronautics and Space Administration (NASA)'s Detect and AvoID Alerting Logic for Unmanned Systems (DAIDALUS), General Atomics Aeronautical Systems, Inc. (GA-ASI)'s Conflict Prediction and Display System (CPDS), and the Airborne Collision Avoidance System-Xu (ACAS-Xu) algorithms. This paper evaluates CPDS and ACAS Xu using the performance metrics explained in RTCA DO-365, and the observed strengths and shortcomings are identified and summarized. The data comes from NASA flight tests (FTs) 2 and 4; which occurred in 2017 and 2016, respectively. The former tested a Phase 2 DAA implementation using ACAS Xu run 3, while the latter tested a Phase 1 implementation using CPDS to perform the RWC function, and TCAS II v7.1 to fulfill the Collision Avoidance (CA) function. In summary, while both algorithms issued the majority of alerts within the MOPS requirements, CPDS tended to out-perform ACAS Xu. CPDS had fewer early alerts and late alerts, indicating that ACAS Xu may need adjustments to tune for MOPS compliance.