Sonar-based local path planning for an AUV in large-scale underwater environments

Abstract

A real-time local path planning for an Autonomous Underwater Vehicle (AUV) in a completely unknown underwater environment is addressed in this paper. Firstly, image processing techniques are applied to extract information of obstacles from the sonar data to build a grid map of the local underwater environment around the AUV. Then a topologically organized bio-inspired neurodynamics model based on the map is constructed to represent the dynamic environment. The goal globally attracts the AUV through the dynamic neural activity landscape of the model, while the obstacles locally push the AUV away to avoid collisions. Finally, the AUV finds its path to the goal autonomously by a steepest gradient ascent rule. The proposed model does not require any prior knowledge of the environment or learning procedures. The simulation result shows that the proposed approaches are capable of planning collision-free paths in large-scale unknown underwater environments.