Distributed Optimal Channel Access in Cognitive Radio Networks

Cognitive radio networks (CRNs) let unlicensed or secondary users (SUs) access licensed spectra, improving wireless resource utilization and addressing the problem of spectrum shortage. SUs are allowed spectrum access only when they don’t cause unacceptable levels of interference to licensed or primary users (PUs). Although throughput-optimal algorithms based on the well-known maximal-weight scheduling algorithm exist for CRNs, they require central processing of network-wide SU information. Developing a distributed implementation that can fully leverage the spectrum opportunities for SUs has so far remained elusive. Because distributed algorithms eliminate the need for network-wide state information collection and avoid costly algorithm executions, researchers Shuang Li, Eylem Ekici, and Ness Shroff developed a fully distributed channel-access solution that uses only local queue-length information. In their article “ThroughputOptimal Queue Length–Based CSMA/ CA Algorithm for Cognitive Radio Networks” (IEEE Trans. Mobile Computing, vol. 14, no. 5, 2015, pp. 1098—1108), the authors identify a new class of queue length–based carrier sense multiple access (CSMA) algorithms as a promising approach to solving the distributed and optimal channelaccess problems in CRNs. This new class of random-access algorithms combines channel-sensing results with locally available queue-length information to determine channelaccess probabilities, achieving full capacity in ad hoc wireless networks in a distributed manner. However, the throughput optimality is realized for nonfading and always-available channel cases, a condition not met in CRNs due to PU activity. Li, Ekici, and Shroff overcome this problem with a novel algorithm based on a new system-state representation that tracks the channel-availability information. This new definition of the system state, along with the differentiated treatment of SUs inside and outside the PU interference range, leads to a distributed algorithm design that achieves provable throughput optimality for SUs without disrupting PU operation. Significant capacity gains realized by the algorithm vis-à-vis other queue length–based CSMA algorithms highlight the efficacy of the authors’ design approach.