Power Control for MACA-based Underwater MAC Protocol: A Q-Learning Approach

Underwater acoustic sensors are battery-powered and spend a major portion of their limited energy during packet transmissions. To conserve energy, multiple access collision avoidance (MACA)-based MAC protocols are designed to lower the data packet transmission power, while using the maximum transmission power for control packets. However, lowering the data transmission power make the data packets susceptible to collisions. In this regard, a reinforcement learning-based power control scheme is proposed for MACA-based underwater MAC protocol that can reduce collisions while maintaining high energy efficiency. A key feature of the proposed scheme is that it enables the sensor nodes to prevent collisions without any prior knowledge of the interferences, eliminating the need for additional signaling. Simulation results show that the proposed scheme significantly improves the energy efficiency and the throughput of MACA-based power control schemes.