Traffic-Aware Optimal Spectral Access in Wireless Powered Cognitive Radio Networks

Traffic patterns associated with different primary users (PUs) might provide different spectral access and energy harvesting opportunities to secondary users (SUs) in wireless powered cognitive radio networks (WP-CRNs). Since the traffic applications have their own distinctive patterns, spectral access and energy harvesting opportunities are also expected to be distinctive. In this paper, we propose a novel approach to identify the PU traffic patterns and estimate the energy harvested from each traffic pattern so that SU can maximize its capacity accordingly. More specifically, we propose a theoretical framework based on a variational inference algorithm to cluster various traffic patterns and design a threshold-based SU transmission strategy by taking into account the spectral access and energy harvesting opportunities for each traffic pattern, so as to optimize SU transmission. Through simulations, we demonstrate the effectiveness of the proposed scheme in terms of throughput gains and show the transmission thresholds under various traffic applications (patterns). Further, we illustrate the effects of different collision costs on throughput for different traffic applications using real wireless traces.