This paper proposes a real-time collision avoidance algorithm for midair unmanned aerial vehicles (UAVs). The algorithm is developed based on the geometry of the closest-point-of-approach (CPA) and a collision detection/resolution scheme. To deal with multiple aircraft, the avoidance velocity of one's own aircraft is expressed in an explicit form, where the avoidance velocity is a component of the safety boundary. The safety boundary is projected on the domain of the flight path angles, and sampling points are chosen along the boundary. The domain of the flight path angles provides directions along which the aircraft can move without collisions, which can reduce the computational load for real-time applications. A collision avoidance solution is obtained by examining possible collision cases involving intruders. The singularity of the geometric approach is also analyzed, and an additional boundary for the singularity is adopted to prevent unexpected maneuvers. The validity of the algorithm is demonstrated through numerical simulations for the non-cooperative and multiple aircraft avoidance problems.
