Real-Time 3D LiDAR-Based Obstacle Avoidance for UAV Using Sector-Filtering Algorithm

초록

Unmanned aerial vehicle (UAV) gained a lot of attention because of its maneuverability in 3-dimensional (3D) and resulting in faster travel time to a certain destination. Therefore, UAV can be integrated to various human needs, especially in delivery service, search and rescue (SAR) mission, and utilized as military defense system. However, the fundamental problem of UAV is the need of proper navigation and obstacle avoidance to reach a destination safely. Diverse research of integrating 3D LiDAR for UAV navigation has been done to reduce complexity due to the high number of point cloud. This paper proposed a sector-filtering algorithm to improve existing real-time obstacle avoidance of UAVs using 3D LiDAR. The current proposed sector-filtering is proposed to take the advantages of 360° field of view (FoV) of 3D LiDAR, without sacrificing the computational resources. In addition, this paper implements 3D LiDAR avoidance with real-world experiment mounted on UAV. The results show that the obstacles have been successfully detected, and the avoidance maneuver is performed for the UAV in real time.

키워드