Asynchronous Dynamic Spectrum Access

Abstract

Synchronization among primary users (PUs) and secondary users (SUs) is either explicitly or implicitly assumed in the literature to allow reliable dynamic spectrum access (DSA) and cognitive radio (CR) networking. However, in realistic DSA that supports ad-hoc transmissions, perfect synchronization is not attainable in the absence of a centralized control mechanism, particularly among SUs. Current efforts either ideally assume perfect synchronization among PUs and SUs or consider medium access that does not require network synchronization based on a single SU. In this paper, we propose an implementable asynchronous DSA scheme by practically considering the timing misalignment with multiple SUs competing for the spectrum opportunity. We further employ a game theoretical framework in asynchronous DSA to thoroughly predict the competition among SUs. Based on the equilibrium concept, we identify the feasible operating region, achieving optimal equilibrium that is not only reaching a win-win situation in the sense of PU throughput and spectrum utilization maximization but is robust to potential selfish behavior of SUs as well. Both low-complexity one-shot and sequential-decision access schemes are analyzed. We demonstrate that, when operating in the proposed feasible region, contrary to intuitive conjecture, the asynchronous DSA achieves surprising improvement in spectrum utilization.