Conflict Avoidance Using an Artificial Potential Field and the mCOWEX Algorithm

This paper presents a combination of an artificial potential field (APF) and the modified Constrained Wavefront Expansion (mCOWEX) algorithm for conflict avoidance in UAVs. As the goal of highly automated UAVs is to be used in shared airspace, i.e. airspace that is used by manned and unmanned aircraft, automatic conflict detection and avoidance is a key requirement. The mCOWEX algorithm presents a capable algorithm to avoid complex conflict scenarios but often generates paths that are hard to predict for human pilots. The APF-mCOWEX algorithm presented in this paper generates paths that are more predictable for human pilots while still being able to solve complex scenarios. The APF-mCOWEX algorithm is a modified form of the original mCOWEX algorithm. These modifications include changes to the placement logic of waves in the mCOWEX algorithm and a changed cost function based on APFs. The final algorithm is then validated on different scenarios of varying complexity.