Distributed unmanned aircraft collision avoidance using limit cycle

This paper presents a guidance algorithm for the distributed collision avoidance of multiple unmanned aircraft based on a limit cycle generated around the collision threat. Distributed collision avoidance, on which individual aircraft detects a collision threat and solves it without any explicit communication with the other aircrafts, among multiple unmanned aircraft is performed using the geometric and kinematic information of the existing aircrafts in a discoverable range. The algorithm analyzes the given information by deriving properties such as line-of-sight(LOS) vector, the closest point of approach(CPA), then figures out a collision threat if exists. A limit cycle, which plays a role of a guidance vector field and actually guides the aircraft to maintain predescribed safety radius between the threat and the aircraft, is applied to a stationary and moving obstacles to provide consecutive flight path angles. By doing this, the algorithm eventually guides the aircraft to a collision-free route to accomplish the goal of the given missions. All of these procedures should be processed independently in real-time. Several scenarios are considered for numerical simulations, and the simulation results demonstrate the effectiveness of the collision avoidance performance and behavioral characteristics of the guidance algorithm. Finally, issues including effectiveness and feasibility of the proposed algorithm are addressed.