Coding rate optimization for efficient underwater optical wireless communication

In this paper, we analyze the underwater environmental factors that determine the quality of the underwater channel and implement a Deep reinforcement learning (DRL) algorithm to adjust the coding rate using sensor information attached to the marine surface vehicle (MSV) to improve communication performance when establishing an underwater optical wireless communication (UOWC) link between the underwater sensor node located on the seabed and the MSV on the sea surface. The agent of the DRL model collects turbidity data in real-time and determines the number of repetitions of transmitted data to meet the required packet error rate (PER). To analyze the performance of the proposed algorithm, simulations were conducted in a virtual environment and a water tank, and both experiments achieved the required communication performance.