Real-time implementation of an online path replanner for an AUV operating in a dynamic and unexplored environment

This study describes the implementation of an online path planner in an autonomous underwater vehicle (AUV) system by using an open-source system architecture, MOOS-IvP. The path planner employed a path replanning scheme and the selective differential evolution quantum-behaved particle swarm optimization (SDEQPSO) algorithm. The implementation was based on a modular framework to ensure the robustness of the path replanner during a mission. The performance of the path replanner was evaluated and verified under stochastic processes in hardware-in-the-loop (HIL) tests, in which the replanner interacted with the onboard controllers and actuators of an Explorer AUV. The experimental results showed the path replanner can be run seamlessly with the hardware onboard an Explorer AUV in real time to generate and continuously refine a safe and feasible path for a dynamic and unexplored environment.