Circular Buffered Hash Based Voxel Mapping and Autonomous Flights for Small-sized Drones in GNSS-denied and Cluttered Environments

In this paper, we describe the results of circular buffered hash based voxel mapping and its autonomous flight applications for small-sized drones in GNSS-denied and cluttered environments. The proposed circular buffered hash based voxel data structure can be utilized in various algorithms handling voxel data, and has a favorable characteristic of automatically maintaining a maximum number of most recent voxel data, thereby it prevents a continuous increase in memory usage. We validated practical applicability of the proposed voxel data structure using a small-sized racing drone platform equipped with the Livox Mid-360 LiDAR, confirming the capability of achieving precise autonomous flights in GNSS-denied and cluttered environments.