Table 5.

Summary of the applications of ML for AUV-aided communication.

References	ML Model Architecture	Contributions	Remarks
[109]	DRL algorithm	AUV optimal trajectory	Robust and effective in different kinds of trajectory tracking
[110]	AMPPO-PP and AMPPO-TT algorithms	Autonomous planning, tracking, and emergency obstacle avoidance	Outperformed the classical path-planning algorithm and advanced sampling-based path-planning algorithm
[111]	RL-based methods	Tracking and obstacle avoidance	Effective in completing the tracking task by avoiding obstacles
[112]	CRNN algorithm	Obstacle avoidance	Avoided obstacles with fewer parameters and shorter computation times, provided shorter paths, and improved energy efficiency