Joint Energy and Reliability Optimization in Hybrid Optical Acoustic Underwater IoT Networks

The integration of high-speed short-range optical
communication with long-range acoustic communication has
become a viable option to solve the underwater Internet of
Things (IoT) communication network. Nevertheless, in
dynamic underwater conditions, there is a great challenge in
ensuring energy efficiency and reliable communication at the
same time. The paper introduces a hybrid optical underwater
IoT network based on deep learning and reinforcement
learning that optimizes the energy consumption and
transmission reliability jointly in hybrid networks to achieve
the goals of deep learning and reinforcement learning. An
adaptive control strategy is built on learning to dynamically
negotiate transmission power and frequency as a result of
real-time channel state feedback, whereas a reinforcement
learning agent chooses appropriate modulation and error
control strategies in order to reduce transmission errors in
changing signal-to-noise. The suggested structure allows
coordination in decision-making among layers of
communication, which leads to a consistent performance in
the underwater space with variable conditions. Widespread
simulation experiments prove that the suggested strategy
yields substantial energy usage savings and still ensures low
bit error rates, better signal quality, quicker adaptation, and
enhanced throughput than the traditional underwater routing
and energy management methods. The findings support the
usefulness of having learning-based joint optimization to
enable reliable, energy-efficient communication over the
underwater IoT.