الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 75 |  |  |
| | | from | 75 | | |
Abstract
The cognitive radio CR technology aims to utilize the spectrum by providing licensed spectrum to be accessed by unlicensed users called secondary users SU. This technology assumes that SUs do not affect the licensed users or primary users PUs; the SUs transmit over the unused channels till the PUs reconnect these channels again then SUs evacuate the spectrum immediately. This requires CR network to have four functionalities to achieve this access without interfering PUs. The first is spectrum sensing to sense the presence of PUs providing information about spectrum holes to SUs and the PUs appearance which interrupts SU connection. The second and third functionalities are spectrum decision and spectrum sharing, which perform quality-of-service QoS requirements evaluation and determine channels to be utilized considering fair sharing of the available channels. These functionalities represent the channel assignment technique achieving better utilization of available resources with fair sharing. Channel assignment techniques are our point of research in this thesis. The fourth functionality is spectrum handoff ensuring that interrupted SU will complete his transmission. Firstly, in this thesis, we used block matrices and Kronecker product to implement a previous model which introduced a novel assignment technique. Moreover, we extended this model by applying an automatic repeat request ARQ as an error correction technique. We compared the basic model with the extended one and it is found that, in the case of high packet error rate PER, ARQ is useful to maintain higher average user throughput. Secondly, we proposed two new channel assignment techniques, namely, maximum-utilization assignment technique and multi-channel assignment technique. We introduced analytical framework to evaluate the performance of the proposed techniques using discrete time Markov chain.These assignment techniques achieved higher average user throughput and lower average queue length especially with high arrival rates compared to the basic model.Finally, the proposed techniques were also extended with ARQ as an error correction technique to get the advantage of higher average user throughput in case of high PER.