The Reinforcement-Learning-Based Adaptive Neighbor Discovery Algorithm for Data Transmission Enabled IoT

The Doppler effect, severe multipath, as well as fluctuating impulse response all contribute to the unfavorable temporal unpredictability of the passageway circumstances in underwater systems. In these cases, energy efficiency is of utmost importance since nodes' battery are permanent and are unable to be swapped out. Therefore, in this study, we provide a reinforcement learning strategy for submerged multichip networks of communication with the goal of extending network lifespan and improving the efficiency of underwater channel broadcasts. The relay device may choose the most effective next hop node to transfer data toward a distant gateway device, which can then connect to the terrestrial internet, thanks to the proposed method's use of a Markov undersea channels theory to describe the link state. It is shown via simulation results that the suggested strategy reduces energy usage and improves latency effectiveness, extending the lifespan of the system.