A Graph-Learning-Inspired Resource Optimization for Digital-Twin-Empowered Wireless Body Area Networks

This letter investigates the resource allocation for digital-twin-empowered wireless body area networks (DT-WBANs). We consider a scenario in which multiple heterogeneous invasive and/or noninvasive biosensors are deployed to constantly monitor the patient's physiological signals and wirelessly communicating with a central hub to reliably transmit the sensory data in real time to a paramedic. Specifically, in order to maximize the energy efficiency of the biosensors guaranteeing the reliability of the emergency-critical physiological data transmission and modeling the nondeterminacy of the stochastic environment, we transform the problem into a Markov decision process (MDP) by jointly considering the transmission mode, relay selection, transmission power, and time slot allocation. Afterwards, a random graph-inspired deep deterministic policy gradient algorithm is proposed to solve the problem and reduce the computational complexity. Simulation results verified the feasibility of the proposed scheme.