Sea environment features-reinforced three-dimensional path planning using modified alpha evolution for long-range AUVs

Three-dimensional (3D) path planning is of importance for long-range AUVs, but suffers from the balance between optimization performances and computation costs. A sea environment features a reinforced path planning method that is proposed using modified alpha evolution (SEF-AE) along with physical constraints. The proposed approach can search for significant features in a 3D sea current environment in the formulation of a hierarchical optimization architecture. Firstly, the vortex-aware reference point sampling is used to construct a navigable region. Then the goal-directed reference path is generated via the second path planning. Finally, the modified AE is used to achieve multi-objective path refinement. A series of experiments referring to the specifications of the Explorer 1000 long-range AUV were carried out with a comprehensive discussion in terms of computational cost, solution accuracy, and robustness. Rigorous comparative results show that the proposed SEF-AE generates lower total-cost paths with the best performance in temporal efficiency, safety compliance, and energy consumption.