Body-OGM Based Grid Mapping Algorithm for Obstacle Avoidance of UAVs

The real-time establishment of obstacle avoidance mapping and path planning determines the smoothness of flight trajectories. Currently, grid mapping entails significant resource and time. However, only perception data within the UAV's vicinity is utilized in obstacle avoidance path planning, grid mapping calculation process involves a significant amount of redundancy. In order to address the issues of lengthy establishment of UAV obstacle avoidance maps, based on the body scope, a new dynamic strategy for creating an Occupancy Grid Map is proposed. Pre-processing lidar point clouds and, determining grids occupancy states through ray-cast and Hit-Miss are both applied to reassign the mapping task and optimize construction cost in multi-core environment. Experimental results demonstrate that the proposed method reduces construction time of occupied grid maps, compared with the original algorithm, and satisfies the requirement of trajectory planning in a dynamic environment.