A parallel genetic approach to path-planning with upstream-current avoidance for multi-AUG deployment

Autonomous underwater gliders are robotic underwater vehicles that do not require operator input. They are frequently deployed in large-scale, long-term ocean-sampling missions that take advantage of their buoyancy-driven engines and very low power consumption. A limitation of the buoyancy-driven engine design is that the glider must travel at slow speeds and is unable to confront strong upstream currents. Optimal path-planning to minimize the upstream-current effect and distance traversed is helpful to guide the glider navigating in the ocean. In a glider path-planning problem, reachability signifies that a minimal upstream-current effect along a path is achieved, while efficiency refers to optimizing path distance, to make it as short as possible. Reachability and efficiency are sometimes inconsistent goals in glider path-planning, and obtaining an optimal solution between both aspects of path-planning constitutes a multi-objective optimization problem. In order to discover an optimal path for glider safety, a parallel genetic approach to the glider path-planning is developed here. We present a novel scheme of upstream-current avoidance to solve the critical path-planning problem in deployment of multiple gliders. The benefits of the proposed approach are used to produce characteristic curves that express how improved cruising capability benefits path diversity. The reachability of an important scheme of upstream-current avoidance is validated by a maximum likelihood function. Moreover, a new crossover operator is proposed to encourage individuals of offspring in solution diversity. Numeric results demonstrate that the proposed path-planning approach determines an optimal path that reduces the upstream-current effect, and also shortens the cruising distance in multi-glider path-planning solutions.