An Investigative Study Into An Autonomous UAS Traffic Management System For Congested Airspace Safety

• Main Authors: Watkins, Lanier, Sarfaraz, Nick, Zanlongo, Sebastian, Silbermann, Joshua, Young, Tyler, Sleight, Randall
• Format: Conference Proceeding
• Language: English
• Subjects: autonomy; Conferences; drone; FAA; Path planning; Planning; Production; Prototypes; Safety; UAS; UTM
• ISSN: 2694-2941
• DOI: 10.1109/ICCWorkshops50388.2021.9473838

Currently, Unmanned Aircraft System (UAS) Traffic Management (UTM) is the Federal Aviation Administration's (FAA) vision for air traffic management below 400 feet. Production UTM systems tend to reside only at specialized test sites and operational centers. UTM has been articulated as a concept of operation (ConOps) by the FAA. The UTM ConOps describes a complex interaction between UAS, UAS Operators, and the UTM system itself. These interactions may involve human operators, or be fully automated. Currently, most UTM studies and experimental prototypes do not look at the UTM concept from end-to-end; instead, they focus on specific aspects of UTM and thus cannot explore and test the holistic performance of a UTM ecosystem. Equally important is ensuring that production UTM can scale to meet the demands of future airspace, which is estimated to be 65,000 UAS operations (takeoffs and landings) per hour by 2035. The busiest US airports currently handle 300 operations per hour.In this paper, we evaluate a portion of the UTM system using a set of autonomous algorithms for flight plan de-confliction. Preliminary results suggest that the autonomy algorithms used for path planning, strategic de-confliction, and detect and avoid (DAA) are capable of scaling to high-congestion scenarios while drastically reducing collisions between UAS, even with almost all UAS deviating from de-conflicted plans (i.e., rogue UAS). We also observed that de-confliction algorithms represent a dominating safety layer in the separation hierarchy, since the strategic de-confliction algorithms manage airspace density, albeit at the cost of longer mission completion times. Our testing was done using a MATLAB simulator, which used the RRT* algorithm for flight planning, two different schedulers (Genetic Algorithm and the NASA Stratway Strategic Conflict Resolution algorithm) for strategic de-confliction, and the Airborne Collision Avoidance System for small unmanned aircraft systems (ACAS sXu) for DAA.