3D Drone Path Planning Algorithms with Collision Avoidance: Collision Resolution vs. Prevention

Because of their limited flight range, multiple drones are often deployed simultaneously to perform complex tasks. The flight path is planned for each drone to follow to complete the job before task execution. However, multi-drone path planning places drones at risk of in-flight collisions. To overcome this problem, we model the multi-drone path planning problem as a multi-vehicle routing problem that maximizes job coverage subject to collision-free paths. We propose three 3D collision-free path planning algorithms, namely, XTRACT, 3DETACH, and ASCEND. XTRACT and 3DETACH provide collision-free paths by setting partial paths at different altitudes while ASCEND prevents intersecting paths at the planning phase by selecting different altitudes. We limit the search processes to two altitudes to demonstrate sufficiency with the lowest height complexity. Through exhaustive evaluations, we compare the performance of the proposed schemes and show the trade-offs between resolving and preventing collisions from path planning. We identify the best-performing strategy by using a profit model to evaluate the plethora of applicable performance metrics.