Selective Velocity Obstacle Method for Deconflicting Maneuvers Applied to Unmanned Aerial Vehicles

摘要

This Note elaborates the novel selective velocity obstacle method that supports unmanned aerial vehicle autonomous collision avoidance. The selective velocity obstacle (SVO) is shown to be able to generate deconflict maneuvers separately to resolve heterogeneous situations, while obeying the right-of-way rules, using a distance-based prescribed turning rate. Problems in the original velocity obstacle method are resolved in the SVO method and validated trough a Monte Carlo simulations procedure. The procedure also reveals that, in a random situation setup, multiple encounters are very unlikely to happen; hence, the SVO method only needs to resolve conflicts in one-by-one manner. Several challenges in the method are noted, such as the late onset in restoring to the mission mode. The method also does not consider a wide range of vehicle types, possible uncertainties in the exchanged data, and three-dimensional conflicts. The results, however, show that the SVO method is a suitable method to reduce the risk of collision, and hence increase the safety of unmanned aerial vehicle operations in the airspace system.