Efficient time-optimal path planning of AUV under the ocean currents based on graph and clustering strategy

Abstract

Path planning is particularly important for Autonomous Underwater Vehicles (AUVs) in the marine environments. Due to the existence of flow fields, many conventional path planning methods are not effective. As the flow distributions vary slowly in large spatial scales, this paper considers the influence of static non-uniform vector field and studies the time-optimal path planning of AUV. The method proposed in this paper is called GK-OPM (Graph and K-means based Optimization Method). GK-OPM generates initial paths for minimum-time path optimization using digraph and K-means method, which benefit optimality and efficiency, respectively. To discuss the parameters in GK-OPM and compare the performance with genetic algorithm (GA) and particle swarm optimization (PSO), Monte Carlo simulation is carried out under the ocean currents simulated by the Lamb-vertex function. The feasibility is verified on discrete ocean data and three-dimensional environments, too. The results show that GK-OPM has about 6–8 times better efficiency than GA and PSO and maintains optimality at the same time.