Age-of-Information Minimization in Healthcare IoT Using Distributionally Robust Optimization

Abstract

In this article, we consider a cellular-based healthcare Internet of Things (IoT) system with imperfect channel state information (CSI), where a healthcare IoT device first receives radio frequency (RF) energy from the small cell base station (SBS) and then transmits physiological status updates to the corresponding SBS as timely as possible. A newly proposed metric, named Age of Information (AoI), will be introduced to characterize the data freshness, which is determined by the status updates generation probability and the information transmission outage probability. To minimize the average AoI, we formulate a distributionally robust optimization problem under an energy harvesting probability (chance) constraint and an information transmission probability constraint. Since the distributionally robust probability constraints are nonconvex, we use the conditional value-at-risk (CVaR)-based method to express constraint specifications related to distributional ambiguity. To tackle the NP-hard problem efficiently, we decompose the AoI minimization problem into two subproblems and propose a low-complexity iterative algorithm to obtain a suboptimal solution. Simulation results show that there exists an AoI-energy tradeoff in the considered healthcare IoT, and the CVaR-based method can achieve a better performance than the nonrobust method.