Cognitive Multi-Channel MAC Protocols with Perfect and Imperfect Sensing David Tung Chong Wong, Shoukang Zheng and Ying-Chang Liang Institute for Infocomm Research, A*STAR, Singapore 1 Fusionopolis Way, #21-01 Connexis, Singapore, 138632 Email: {wongtc, skzheng, ycliang}@i2r.a-star.edu.sg Abstract—Analytical formulations of the throughput of cogni- tive multi-channel MACs with perfect and imperfect sensing are presented. Both imperfect concurrent sensing and imperfect se- quential sensing schemes are considered. A discrete time Markov chain is used to model the number of communicating node pairs in the MAC protocols. The throughput of the MAC protocol with perfect sensing is expressed as a function of the number of available data channels, the channel transmission rate, the average utilization per channel, the steady state probability of having a number of available data channels, while the throughput of the MAC protocol with imperfect sensing is expressed as a function of the number of available data channels, the channel transmission rate, the average utilization per channel, the steady state probability of having a number of available data channels, probability of false alarm and probability of misdetection. The results also clearly demonstrate the advantage of our proposed MAC protocol with imperfect concurrent sensing having low probability of misdetection but high probability of false alarm. Index Terms—MAC protocols, perfect sensing, imperfect sens- ing, probability of misdetection, probability of false alarm. I. I NTRODUCTION Cognitive radio is a hot research area in recent years. IEEE 802.22 draft standard is using cognitive radio in wireless regional area network (WRAN), while IEEE 802.11af draft standard is looking at operating WiFi in TV white space. Medium access control (MAC) protocols with cognitive radio capabilities are also proposed [1],[2]. Reference [1] considers a cognitive IEEE 802.11 MAC, while reference [2] considers multiple cognitive CSMA/CA networks that can coexist to- gether. Comparison of multi-channel MAC protocols is also studied in [3]. Reference [3] compares the throughput perfor- mance of a number of multi-channel MAC protocols, including Dedicated Control Channel, Common Hopping, Split Phase and Parallel Rendezvous protocols. However, cognitive radio is not considered. Reference [4] considers a popular scenario of a primary network and a secondary network where the primary network has priority for the usage of the spectrum band over the secondary network. Reference [5] extends multi-channel MAC protocols for opportunistic spectrum access (OSA). The design of multi-channel MAC protocols for OSA in ad hoc networks is considered in [6]. The modeling in [6] embeds the probability of detection and probability of false alarm in their transition probability, modeled by a discrete time Markov chain. These probabilities are not expressed explicitly in the analytical throughput expression. In our paper, we consider cognitive multi-channel MAC protocols using Dedicated Control Channel with perfect and imperfect sensing. One of the channels is used as the control channel with request-to-send (RTS) and clear-to-send (CTS) messages, while the rest of the channels are used for data packet transmissions if the channel is not occupied by a primary user (PU). The activity of a PU is modeled by a two-state discrete time Markov chain. The contribution of this paper is as follows. First, a discrete time Markov chain is used to model the number of communicating node pairs in the MAC protocols with the probability that at least one channel is detected as available out of the remaining channels or remaining number of pairs of communicating nodes explicitly embedded in the transition probability. Second, the throughput of the MAC protocol with perfect sensing is expressed as a function of the number of available data channels, the channel transmission rate, the average utilization per channel, the steady state probability of having a number of available data channels, while the throughput of the MAC protocol with imperfect sensing is expressed as a function of the number of available data channels, the channel transmission rate, the average utilization per channel, the steady state probability of having a number of available data channels, probability of false alarm and probability of misdetection. Finally, the results in the numerical section demonstrated the advantage of our proposed MAC protocol with imperfect concurrent sensing having low probability of misdetection but high probability of false alarm. The rest of the paper is organized as follows. Section II describes our cognitive multi-channel MAC protocols. In Section III, we present analytical models for our proposed MAC protocols. Numerical results are presented in Section IV. Both analytical and simulation results are presented. Finally, concluding remarks are made in Section V. II. MAC PROTOCOLS We present MAC protocols with perfect and imperfect sensing. Let M denote the number of channels and M D the number of data channels available, excluding the channels used by the communicating secondary nodes and the channels detected as active. A. Perfect Sensing Multi-channel MAC transmits in a number of channels at the same time, resulting in increase in throughput. We consider a multi-channel MAC using a dedicated control channel and M - 1 data channels with a PU in each of the data channels. 978-1-61284-231-8/11/$26.00 ©2011 IEEE This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2011 proceedings