Back-Pressure Based Throughput Enhancement Algorithms for Cognitive Radio Networks
Open AccessIn the recent years, with growing demands on wireless communications, techniques to efficiently utilize the limited wireless resources become vital. Wireless spectrums are managed and allocated to licensed users by the Federal Communications Commission (FCC). However, the total amount of available spectrums is limited due to the physical constraints. On the other hand, many allocated spectrums are not fully utilized. Cognitive Radio Networks (CRNs) technique allows unlicensed users (Secondary Users (SUs)) to opportunistically access the licensed spectrum without interfering with licensed users (Primary Users (PUs)), to exploit the under-utilized portion of the spectrum. Due to the characteristics of CRNs, such as the coexistence of PUs and SUs and the highly dynamical network topology, most of current scheduling algorithms designed for Wireless Mesh Networks (WMNs) cannot be adopted directly to this field. The back-pressure scheduling framework, for instance, can guarantee the throughput to be optimized for a range of wireless networks. However, the original back-pressure focuses on a single channel with fixed connectivity and thus may not be feasible to the CRN. In addition, the optimization problem contained in the original back-pressure approach requires significant computation time, which is not feasible in a dynamic environment such as CRNs.In this research, we focus on developing back-pressure based scheduling algorithms for CRNs, and make effort to design different approaches aiming at different objectives. Our research starts from investigating the medium access control method for throughput enhancement in an 802.11 wireless network. We show that it is possible to achieve a global objective such as aggregated throughput enhancement and fair sharing the channel occupancy time, by distributively controlling the behavior of each station in a network. We then analyze the relationship between the transmission probabilities of SUs and the stability of the network, and investigate the sufficient condition for achieving throughput optimality. The results can be later adopted to design our distributed scheduling algorithm which is suitable for a CRN with dynamical routing. Furthermore, we revise the backlog function to find a delay-aware approach, which can overcome the potential long delay problem in the original scheme. The throughput optimality property of the revised backlog is analyzed under a CRN scenario, and a sub-optimal greedy solution is proposed in this work.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.