Adaptive Transmission Scheduling in Time-Varying Underwater Acoustic Channels

We consider an underwater acoustic communication system where each packet arrives periodically at the transmitter, and is required to be sent to the receiver under an individual and predetermined delay constraint. Given the high temporal dynamics of underwater acoustic channels, it is desirable to adapt the transmission schedule to channel conditions for transmission energy efficiency. With a goal of minimizing the total packet transmission energy consumption, and assuming perfect and non-causal channel knowledge, a waterfilling algorithm is developed to recursively determine the optimal transmission schedule of each packet. Given the channel causality in practical systems, we focus on a particular type of underwater acoustic channels that exhibit periodic dynamics, and apply the Holt- Winters approach for channel prediction. An online scheduling algorithm is designed to decide the optimal action of each time slot (i.e., to transmit or hold the packet on the top of the transmitter packet queue) based on the predicted channel condition in future time slots and the packet queue status. The proposed algorithms are evaluated in simulated channels and using channel measurements from field experiments. Numerical results demonstrate that the proposed algorithms yield considerable transmission energy saving relative to a benchmark method that transmits each packet upon its arrival.