Global path planning and waypoint following for heterogeneous unmanned surface vehicles assisting inland water monitoring

The idea of dispatching multiple unmanned surface vehicles (USVs) to undertake marine missions has ignited a burgeoning enthusiasm on a global scale. Embarking on a quest to facilitate inland water monitoring, this paper presents a systematical approach concerning global path planning and path following for heterogeneous USVs. Specifically, by capturing the heterogeneous nature, an extended multiple travelling salesman problem (EMTSP) model, which seamlessly bridges the gap between various disparate constraints and optimization objectives, is formulated for the first time. Then, a novel Greedy Partheno Genetic Algorithm (GPGA) is devised to consistently address the problem from two aspects: (1) Incorporating the greedy randomized initialization and local exploration strategy, GPGA merits strong global and local searching ability, providing high-quality solutions for EMTSP. (2) A novel mutation strategy which not only inherits all advantages of PGA but also maintains the best individual in the offspring is devised, contributing to the local escaping efficiently. Finally, to track the waypoint permutations generated by GPGA, control input is generated by the nonlinear model predictive controller (NMPC), ensuring the USV corresponds with the reference path and smoothen the motion under constrained dynamics. Simulations and comparisons in various scenarios demonstrated the effectiveness and superiority of the proposed scheme.