Achieving optimal-dynamic path planning for unmanned surface vehicles: A rational multi-objective approach and a sensory-vector re-planner

Abstract

Navigating unmanned surface vehicles (USVs) in the intricate and unpredictable oceanic environment poses an indispensable challenge for path planning, necessitating meticulous strategies and unwavering resolve. In this paper, under complex and unforeseen circumstances, a novel path planning framework incorporating the multi-objective optimization and a sensory-vector replanning strategy is created for a USV. First, by encapsulating the intricate nature of ocean environment and ship dynamics, a nonlinear multi-objective path planning problem is designed, thereby providing a comprehensive and in-depth portrayal of the underlying mechanism. By integrating the principles of candidate set random testing and adaptive crowding distance, an adaptive enhanced non-dominated sorting genetic algorithm (AENSGA-II) is devised to fully exploit the underlying optimization problem in constrained dynamics. To avoid over-subjective choice in the Pareto set, a fuzzy-linguistic satisfactory degree is deliberately designed, where the linguistic importance preference of the objectives is re-evaluated in the Pareto set, facilitating the decision-making. By inserting virtual sensory vector onto the USV, a seamless interface between global planning and COLREG-compliant replanning mechanism is devised, thereby contributing to the entire hierarchical scheme. Eventually, the framework merits autonomous global-planning and local-reaction in an organically way. Comprehensive simulations and comparisons in various ocean scenarios demonstrate the effectiveness and superiority of the proposed path planning framework.