A global path planning method for a net-cleaning underwater vehicle (NCUV) based on hierarchical collaborative optimization and pollution level classification

To meet the industrial demand for cleaning marine aquaculture net cages and to address the low efficiency and ecological impact of current cleaning operations, this paper proposes a path-planning method for a net-cleaning underwater vehicle (NCUV) based on hierarchical cooperative optimization (HCO). Conventional ant colony optimization (ACO) and grey wolf optimizer (GWO) algorithms typically perform optimization only at a single level, and thus struggle to effectively handle redundant inter-region connections and low coverage efficiency within regions. The proposed HCO approach adopts a two-level cooperative mechanism that combines global region sequencing with local path refinement, thereby resolving both issues. At the global level, the traversal order of sub-regions is optimized using the L-SHADE algorithm, whereas at the local level an improved GWO introduces a nonlinear function to control the convergence factor and thereby optimizes the trajectory within each sub-region. Experimental results demonstrate that, in terms of path planning, the HCO method outperforms traditional ACO and GWO algorithms, achieving significant improvements in cleaning efficiency and energy consumption.