| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 105 === With the increasing requirement for wireless communications and the rapid development of communication technologies, the demands for bandwidth and data transmission rate are growing. On the other hand, the available spectrum is not only scarce but also under-utilized. To improve the utilization of radio spectrum, the concept of Cognitive Radio (CR) is proposed. Cognitive radio, which allows secondary users (SUs) to opportunistically access a spectrum hole which is in the same frequency bands as primary users (PUs), provides an efficient approach to increase the availability of idle spectrum to new services. Traditionally, PUs have the preemptive priority over SUs, and hence the acquisition of underutilized channels by SUs is not guaranteed. To improve the quality of service (QoS) of SUs, we consider multi-rate cognitive radio networks (CRNs) with not only a buffer for preempted SUs but also spectrum leasing for all SUs. Three call admission control (CAC) schemes with spectrum leasing are also developed: CAC1, CAC2, and CAC3. Moreover, we also propose two variations of CAC3, i.e., CAC3a and CAC3b. With CAC1, the bandwidth requirement of each SU is one channel, that with CAC2 is a constant bandwidth which is greater than one, and that with CAC3 is a variable bandwidth which is between a lower bound and an upper bound. In addition, spectrum leasing for SUs are enforced, where the licensed spectrum band is divided into two parts, i.e., the leased spectrum and unleased spectrum, and for all CAC schemes PUs only have the preemptive priority over SUs in the unleased spectrum. Furthermore, to mitigate the dropping probability of preempted SUs, a buffer is utilized to accommodate the preempted SUs till their maximum waiting time expires. The analytical models are derived for all considered CAC schemes with or with buffer. We develop an iterative algorithm to find the steady state probability and evaluate the performance measures of interest. The performance measures of interest include the blocking probability, throughput, the average number in the system, and the average system delay for PUs and SUs, respectively, the dropping probability for SUs, the SU cost, and the SU utility. Furthermore, we compare the performance of the system with buffer and that without buffer. We also compare the performance of CAC1, CAC2, and CAC3a (CAC3b). Last but not least, computer simulation is written in Visual Studio 2012 to validate the accuracy of the derived analytical models.
|
|---|
