Research on local path planning for UUV based on improved Q-learning algorithm
Compared to the traditional Q-Learning algorithm, the improved algorithm significantly reduces the number of episodes required to find the optimal path, produces smoother paths closer to the optimal solution, and achieves substantial improvements in convergence speed, exploration efficiency, and path planning quality.