Improving the Quality of Underwater Wireless Optical Communications in Uncertain Ocean Environments

Underwater wireless optical communication can be used between multiple autonomous underwater vehicles to enhance information sharing during ocean exploration. The quality of optical communication is affected by many marine environmental parameters, such as ocean currents, disturbances, and attenuation. This study proposes using a reinforcement learning algorithm and adding random parameters and thruster delay in the simulated environment to improve the stability of the link. The trained policy enhanced the stability of the link in the simulated environment. This method does not need to model the complex and unknown ocean environment and can incorporate more parameters that affect the quality of underwater optical communication in the future.