Collision risk management for non-cooperative UAS traffic in airport-restricted airspace with alert zones based on probabilistic conflict map

Abstract Recent years have seen an increase in reported unmanned aerial systems (UAS) incursion into terminal airspace, likely due to the ease of access to recreational UAS to the general public with minimal understanding of aviation regulations. Such incursions often lead to extensive airport shutdowns due to safety concern and could cause a cascading disruption to airline operations throughout the region. A better assessment tool for the collision risk between the existing air traffic and such non-cooperative intruder could help reduce unnecessary disruption while maintaining safety to air traffic operations. While advancements have been made in the past decade in detection and avoidance of UAS traffics, the systems depend on either having fully cooperative traffic or sensor capability beyond what is currently available. This paper seeks to assess the collision risk posed by an intruding UAS within the airport-restricted “terminal airspace,” in the Singapore context, with minimal information on the UAS. This was done through probabilistic conflict prediction with Monte-Carlo simulations under the assumption of a worst-case scenario involving a non-cooperative intruder traveling directly towards the flight corridor. Alert Zones within the existing 5 km terminal airspace could be constructed using the collision prediction models to help air traffic controllers quickly identifying UAS that poses a threat under most circumstances, with the information passed on to the pilot for mitigation actions. Simulations were conducted for a number of conflict-pairs to investigate how the resulting Alert Zones differ. Finally, this paper also investigates how the incorporation of UAS tracking information, under the current reliability level, could be used to augment the collision prediction algorithm and its effect on the management of collision risks in terminal airspace. The result suggested that the method that utilizes the most available information for UAS path modeling could be used with the currently available sensors to produce acceptable collision prediction results.