A Dual-Frequency Data-Driven Coverage Path Planning Algorithm for Unknown Large-Scale Marine Area

Abstract

Detecting regions in an unknown large-scale marine area by a side scan sonar on an Autonomous Underwater Vehicles (AUV) as quickly as possible is often of great importance. Recently, high-resolution sonar, known as the dual-frequency identification sonar (DIDSON), produced near-video quality images. It has two different operation modes: the detection mode and the identification mode. In this paper, the dual-frequency data-driven coverage path planning (D2-CPP) algorithm is proposed, which provides a coverage solution based on real-time local information gain of DIDSON data in different working modes. Through simulation and field trials, it is proved that the D2-CPP significantly decreases the coverage path length while covering the same area by identification mode as the standard lawnmower.