Optimal Motion Planning for Heterogeneous Multi-USV Systems Using Hexagonal Grid-Based Neural Networks and Parallelogram Law Under Ocean Currents

This article addresses the challenge of enhancing collaboration efficiency within a heterogeneous system of multiple unmanned surface vehicles (USVs) while accounting for the impact of ocean currents. In this context, this article introduces an intelligent algorithm called the hexagonal grid-based neural network with parallelogram law (HGNNPL). The algorithm comprises three key components: 1) a bio-inspired neural network (BINN) designed to predict an optimal collision-free path for a multi-USV system, which operates based on hexagonal partitioning grids, ensuring smooth navigation without collisions; 2) an adjustment component plays a crucial role in correcting deviations caused by ocean currents and calculating the associated energy consumption; and 3) an optimal task assignment component responsible for assigning task objectives to the USVs, where distance determined by the BINN and the energy consumption are involved as motion planning costs. This article presents simulation results that compare the performance of the proposed algorithm with an existing algorithm based on square grids, which does not account for the elimination of ocean current effects. These results illustrate the practical effectiveness of the proposed method.