Parallelization of Local Path Planning for High Reliable Autonomous Drones

Yusuke Hirota, Ittetsu Taniguchi, Takao Onoye

Abstract: Safety is crucial for autonomous drones. This paper proposed a new application to realize a high-reliable drone under the edge computing environment. To realize safer drones, we focus on the unexpected and unsupposed situations in drone flight and aim to cope with the situations by calculating the countermeasures for various expected scenarios beforehand. Especially, this paper focused on local path planning, and parallelization of the local path planning is proposed to calculate the many paths to escape the expected obstacles in various scenarios. Experimental results show the effectiveness of the proposed method.