Obstacle Avoidance with Zigzag Tentacles for Multirotor UAVs
Kazuya Arashi, Naoki Akai, Kane Saliou, Susumu HARA · 2023 IEEE/SICE International Symposium on System Integration (SII) · 2023
In this paper, we present an extended tentacle-based obstacle avoidance method for multirotor unmanned aerial vehicles (UAVs). Our objective is to achieve a rapid and safe autonomous flight in narrow and complex spaces. Traditional tentacle-based obstacle avoidance methods generate only simple-shape candidate paths. Consequently, paths that bypass obstacles cannot be found and the moving speed of a UAV is needed to be decreased when navigating narrow spaces because the tentacles cannot be sufficiently extended. In the presented method, zigzag tentacles are generated. A UAV can fly through narrow spaces without decreasing its speed because the zigzag tentacles can find bypass paths. We conducted experiments with simulated and actual multirotor UAVs. The results show that the presented method achieved a rapid and safe autonomous flight in both experiments. In addition, we compare the presented method with the dynamic window approach which is a traditional tentacle-based method. This comparison shows the effectiveness of the zigzag tentacles, i.e., the presented method can achieve autonomous flight whereas the traditional method causes the UAV to become stuck.