An autonomous integrated system for 3-D underwater terrain map reconstruction

This paper presents a novel integrated approach of creating a 3-D surface map of seabed terrain using an Autonomous Underwater Vehicle (AUV) equipped with various sensors such as multi-beam sonar sensors, DVL, and IMU. The underwater terrain map is useful for various applications like fishery, search and rescue operations, underwater surveillance, mine hunting, etc. The acoustic sensors mounted on the AUV overcome the unfavorable underwater conditions like nonuniform illumination and restricted visibility. In the proposed methodology, the AUV is autonomously navigated using a navigation controller on a flat 2-D surface at a certain depth, to cover an a priori unknown area using an adaptive path planning algorithm. The terrain data is captured by the down-facing multi-beam sonar sensors. The Long Baseline (LBL) localization system and PID controllers are used to drive the AUV to the set-points given by the navigation controller. The collected data is then used offline to construct a 3-D map of seabed using alpha shapes. The performance of this method is tested and validated using a high-fidelity underwater simulator UWSim based on Robot Operating System (ROS). In the simulation runs, the proposed methodology is shown to reconstruct accurate 3-D maps of the seabed.