UAV Collision Avoidance Using RRT* and LOS Maximization

Unmanned Aerial Vehicles (UAVs) are becoming increasingly useful for several different applications. One issue faced in the growing popularity of UAVs is collision avoidance. With multiple aircraft in a finite airspace, some sort of algorithm must be used in order to avoid near misses and collisions while staying within reasonable bounds for a given mission. This paper proposes a new algorithm based on three methods of path finding: Dubins path, line of sight (LOS) maximization and a modified version of rapidly exploring random trees (RRT*). The algorithm is used in a realistic ROS based simulator to obtain results presented in this paper.