Automatic Multiagent Aircraft Collision Avoidance

Abstract

In recent years, there have been incentives to move towards the Free Flight concept in aviation due to advances in technology and increasing air traffic. Free Flight proposes a distributed cooperative strategy based on direct communication between aircraft to maintain separation rather than the currently centralized manner that is handled by the air traffic controllers. Free Flight requires aircraft to be able to detect and resolve conflicts that could lead to loss of separation between multiple aircraft at any point during the flight, which proves to be a challenging task. This research seeks to find an optimized cooperative collision avoidance strategy to resolve conflicts in the horizontal and vertical planes by proposing maneuvers that involve changing the altitude or heading involving multiple aircraft. The method employed is to extend the Zeuthen strategy—an adaptation of the Monotonic Concession Protocol (MCP)—from pairwise negotiations to handle multiway conflicts. This coordinated strategy provides optimal change of trajectories through resolutions that maximize the product of agents’ utilities. We show with a simulator that this distributed approach is efficient and prevents losses of separation even in a highly congested airspace. The simulator also provides an infrastructure, on which alternative algorithms can be empirically evaluated.