Table 11.

Heterogeneous cooperation intelligence algorithms.

Problem	Resolution	Performance and Additional Explanation	Ref.
Heterogeneous system formation (UAV–USV–UUV)	DQN (deep Q-learning) algorithm	LoS (line of sight) (UUV–USV) and underwater acoustic channel (USV–UUV) Markov decision process for control A success rate of target hunting over 95% A joint 3U heterogeneous system Balanced system energy consumption and interconnectivity	[180]
USV–UAV Systems	Multiultrasonic joint dynamic positioning algorithm	Multiultrasonic joint dynamic positioning algorithm G: hierarchical landing guide point generation algorithm and cubic B-spline curves UAV can land on the USV in 10 min The multiultrasonic joint dynamic positioning algorithm is based on ToA, which shows the position of the UAV in real time Cooperation mechanism and motion environment research	[181]
USV–UAV structure	CamShift algorithm and Douglas–Peucker algorithm	Turning mode and PID mode for control Useful for real-life maritime search and lifesaving missions Rescue operation using USV–UAV cooperation Cover and recognize a wider area by inspecting the scene with a UAV USVs bring people to shore, act as buoys, and distribute life jackets.	[182]
UAV–USV–AUV path planning	IPSO (improved particle swarm optimization) algorithm	UAV–USV–AUV systems are more efficient than are USV–AUV systems in performing search and tracking (SAT) missions Study of cooperative path planning problem for search and tracking (SAT) missions for underwater targets using UAV–USV–AUV cooperation The motion of a vehicle is expressed by the equations of motion	[183]