Lyapunov optimization-based power control policy for time-critical applications in a hybrid cognitive radio IoT network

Abstract

Ensuring timely quality of service for critical data in internet of things (IoT) networks is vital, demanding adequate spectrum and energy resources. However, these networks face challenges such as spectrum leakage and the limited power capacity of IoT devices. Merging cognitive radio (CR) technology in these networks is a potential solution to mitigate the challenge of the insufficient frequency spectrum. In this paper, we consider a scenario of two secondary users (SUs) accessing a CR network with a single primary user (PU) within the interweave/underlay access scheme. The SUs have two heterogeneous traffic: status update packets and deadline-constrained packets, which are measured using the age of information (AoI) and the timely throughout metrics, respectively. To address the energy limitation issue, an adaptive power allocation strategy for the SUs’ transmissions is initially developed. Subsequently, we investigate the trade-off between power consumption for the secondary system and specific demands of each traffic type independently. A stochastic optimization problem is then constructed to minimize the weighted sum average power consumption of the secondary system while ensuring a maximum average AoI threshold and an average minimum timely throughput per frame requirement for the corresponding SUs. The Lyapunov optimization theory is employed to convert the formulated problem into an unconstrained Markov decision process (MDP) per frame. Moreover, we derive a formula for the outage probability of SUs’ transmissions, considering the presence of multiple primary receivers. Simulation results demonstrate how the total consumed power is affected by the variations in the key system parameters, such as the activity and transmission power of the PU. Furthermore, experimental simulations show that the proposed policy achieves significant enhancement over two baseline policies.