A novel reinforcement learning based tuna swarm optimization algorithm for autonomous underwater vehicle path planning

Abstract

Path planning technology is an important guarantee for the safe navigation of autonomous underwater vehicle (AUV) in water, and it is also an important indicator of the intelligence of autonomous underwater vehicle. Aiming at the path planning problem of AUV in complex environments, this article presents a reinforcement learning-based tuna swarm optimization algorithm called the QLTSO. In this algorithm, individuals are independent of each other, and the choice of each individual strategy is decided by reinforcement learning. Four strategies are set for each individual in the algorithm: spiral foraging, parabolic foraging, optimization adjustment and ESOS strategy. Finally, the cubic B-spline curve is used to smooth the path so that the autonomous underwater vehicle can better track the path. To verify the superiority of the QLTSO algorithm, the algorithm is compared with other advanced optimization algorithms. The simulation results show that the QLTSO algorithm can plan safe and effective AUV navigation paths in a variety of two-dimensional and three-dimensional complex environments with better convergence and robustness, and the planning success rate is up to 100%, which is an effective AUV path planning algorithm.