Autonomous 3-D mapping and safe-path planning for underwater terrain reconstruction using multi-level coverage trees

Abstract

This paper presents an autonomous approach of 3-D reconstruction of underwater terrain using multi-level coverage trees. An autonomous underwater vehicle (AUV) equipped with multi-beam sonar sensors, Doppler velocity log (DVL) and inertial measurement unit (IMU) sensors is used to achieve this goal. The underwater 3-D search space is represented by a multi-level coverage tree which is generated online based on the obstacle information collected by the AUV. The nodes of the tree correspond to safe sub-areas for AUV navigation which are identified based on obstacle density in neighborhood of free cells. Standard tree traversal strategies like depth-first-search (DFS) and breath-first-search (BFS) are then used for visiting all the nodes of the tree thus securing complete coverage of the 3-D space. The terrain data collected by the AUV during tree coverage is used offline for the 3-D reconstruction of seabed using alpha shapes algorithm. The performance of this method is validated using a high-fidelity underwater simulator UWSim based on Robot Operating System (ROS). The simulations show that the proposed methodology achieves safe path planning and accurate reconstruction of the 3-D map of the underwater terrain.