UAV collision avoidance using multiple artificial potential functions: Practical implementation and experimental outdoor applications

In unmanned aerial vehicle (UAV) applications, fast-response collision avoidance is an essential requirement for a safe flight. This applies in particular to real-world settings with constrained hardware performance. With these challenges in mind, the authors have recently presented a collision avoidance strategy based on multiple artificial potential functions (MAPOFs). The strategy overcomes the typical drawbacks of the conventional, popular APOFs approach, such as chattering or deadlock. The main focus of the present paper is on the practical implementation and experimental demonstration of the proposed MAPOF strategy in two real-life outdoor scenarios using a commercial quadrotor and open-source software. The experiments show the effectiveness of the MAPOFs strategy in the presence of stationary obstacles as well as in the presence of another moving UAV. The experiments are complemented by additional simulation results to further illustrate and validate the proposed approach.