Abstract: To address the issue of low efficiency in large-scale scene exploration when using a fixed-mount LIDAR, an autonomous exploration motion planning method with LIDAR pitch controllable for unmanned aerial vehicles (UAV) is proposed. This method consists of lightweight viewpoint generation and path planning considering characteristics of LIDAR pitch controllable. In the viewpoint generation, a frontier normal-based viewpoint generation method is employed to reduce computational load. Then, the LIDAR pitch is calculated based on the relative position between the frontier and the viewpoint, so as to increase the number of observable frontier at viewpoints. In the path planning, an equivalent viewpoint mechanism considering the characteristics of LIDAR pitch controllable enables adjustments of the LIDAR pitch to reduce UAV yaw adjustment time, which decreases the switch time of trajectory and redundant movements, thereby reducing the overall exploration task duration. Simulation results demonstrate that compared to the FUEL method, the proposed method reduces the average exploration time and path length by about10% in a typical factory scenario, and reduces over 20% exploration time and over 10% path length in a multi-story structured mall.

Key words: unmanned aerial vehicles, autonomous exploration, LIDAR pitch controllable, viewpoint generation, motion planning