الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 135 |  |  |
| | | from | 135 | | |
Abstract
The cognitive radio (CR) system has come into being as a decisive solution to the scarcity and under utilization problems of the spectrum resources. from this point onwards, owing to the CR capabilities, the unlicensed user or so-called secondary user (SU) can access the licensed band without any disturbance to the licensees or so-called primary users (PUs). In such case, the SU supposedly renounces his operating channel upon the PU appearance. Meanwhile, the handoff process is initiated to ensure a seamless transmission of SUs. In this regard, two interesting handoff approaches are suggested by the IEEE 802.22 wireless regional area networks (WRAN) standard, that is, the always-staying and always-changing approaches. The former approach obliges the interrupted SU to constantly wait in silence on his operating channel until being idle again. On the other hand, in the latter approach, the SU is forced to continuously change his operating channel after each interruption. from another perspective, due to the contention between multiple SUs on the same channel, it is worthwhile to consider the QoS requirements of each SU. In most common real life applications, SUs potentially have distinct QoS requirements, upon which they are classified into different classes. Delay-sensitive traffic such as VoIP transmission is given a higher priority over delay-insensitive traffic such as data transmission. Such a situation is mathematically modelled as a priority queueing system. In such a system, the SUs are classically served either in preemptible or non-preemptible manner. It is believed that these classical service disciplines enhance the performance of particular classes at the expense of remarkable degradation in the others. To address this problem, we propose a CR-based hybrid service discipline at which low priority SUs are no longer preempted by high priority SUs when their number of interruptions reaches a certain threshold value. In addition, we propose a special channel access priority to the SUs upon reaching the aforementioned threshold of interruptions in order to further reduce their cumulative delays. In this context, analytical formulas that represent the SUs latency performance and dropping probability are derived based on the residual-life approach of the M/G/1 priority queueing system. This analytical study is firstly applied to the always-staying handoff approach in one typical channel. Based on a set of numerical results presented for a 2-class network, it is shown that selecting an appropriate threshold value yields a more balanced performance that satisfies the needs of all SUs classes without exception. Moreover, the superiority of our model significantly emerges under the heavy loaded traffic of PUs and SUs. The scalability of our work is further investigated by considering the always-changing handoff approach for a CR network with multiple channels. Furthermore, the hybrid service discipline is assumed to be controlled by different threshold values corresponding to each priority class, which is known as class-dependent threshold (CDT) policy. The numerical results for a 4-class network reveal that the CDT approach can provide better performance compared to the unique threshold (UT) policy.