A Complete Coverage Path Planning Approach for an Autonomous Underwater Helicopter in Unknown Environment Based on VFH+ Algorithm

Abstract

An Autonomous Underwater Helicopter (AUH) is a disk-shaped, multi-propelled Autonomous Underwater Vehicle (AUV), which is intended to work autonomously in underwater environments. The near-bottom area sweep in unknown environments is a typical application scenario, in which the complete coverage path planning (CCPP) is essential for AUH. A complete coverage path planning approach for AUH with a single beam echo sounder, including the initial path planning and online local collision avoidance strategy, is proposed. First, the initial path is planned using boustrophedon motion. Based on its mobility, a multi-dimensional obstacle sensing method is designed with a single beam range sonar mounted on the AUH. The VFH+ algorithm is configured for the heading decision-making procedure before encountering obstacles, based on their range information at a fixed position. The online local obstacle avoidance procedure is simulated and analyzed with variations of the desired heading direction and corresponding polar histograms. Finally, several simulation cases are set up, simulated and compared by analyzing the heading decision in front of different obstacle situations. The simulation results demonstrate the feasibility of the complete coverage path planning approach proposed, which proves that AUH completing a full coverage area sweep in unknown environments with a single beam sonar is viable.

Complete Coverage Path Planning Based on Improved Genetic Algorithm for Unmanned Surface Vehicle

“…In addition to the above-mentioned conventional processing methods, some researchers have also tried to use swarm intelligence algorithms to solve the problem. There are artificial potential field methods [24], fuzzy logic algorithms [25], ant colony algorithms [26], genetic algorithms [27], vector field histogram algorithms [28], and biologically inspired algorithms [29]. The advantage of cluster intelligence algorithms is that they have some global optimization capabilities [30], but they often have the disadvantage of slow convergence speed and a large amount of calculation [31].…”