Study on path planning method for unknown 3D environment exploration driven by compound boundary points

Abstract

Aiming at the low exploration level, slow speed and repeated exploration of UAVs in unknown environments, this paper improves the selection of boundary points based on the boundary driving method, and proposes a method suitable for UAVs to quickly explore in unknown environments. Planning method: while using the depth image (RGBD) sensor to receive environmental information, the octree map of the known environment is constructed in real time, and the common indoor environment is divided into six local environmental types. The drone observes according to the RGBD camera on board. The local environment type generates instantaneous speed commands, and designs composite boundary points, from which the point with the largest information gain and the smallest yaw angle is weighed as the boundary guide point. Finally, a simulation experiment was carried out and comparing with the existing methods. The time required to explore the apartment environment in the next best viewpoint (NBV) in Document with the present method is 68.7% less than the exploration time in the paper. In addition, this paper designs There are four types of environments. In these environments, the present method reduces the average exploration time by 97.1% comparing with the method in Document. The experimental results show that the present method can effectively improve the exploration efficiency and has strong feasibility.