Reliable and Energy-Efficient Relay Transmission in WBANs with Wireless Power Transfer: Optimal Design with DRL

Reliable and energy-efficient wireless transmission is critical to the success of future wireless body area networks (WBANs), as the energy storage capacity of biosensors is strictly constrained. In this article, we study the energy efficiency (EE)-maximization problem in a decode-and-forward (DF)-relay transmission system powered by wireless power transfer (WPT). Specifically, in each time slot, an access point (AP) transmits radio frequency (RF) signals to charge a biosensor and a relay, and then the biosensor transmits its collected data to the AP with the assistance of the DF relay. To ensure human safety and maintain transmission reliability, we formulate the EE-maximization problem under peak power constraints while considering transmission errors. Then, we apply a deep reinforcement learning (DRL) algorithm to train an action policy for determining near-optimal transmission parameter values for each transmission session. Through some selected numerical examples, we show that under peak power constraints, our solution can approach the performance obtained by the exhaustive search and exhibits superior performance compared to the traditional iterative algorithm. Therefore, our solution is instrumental in designing safe, reliable, and energy-efficient WBAN systems.