چكيده به لاتين
In cognitive radio networks (CRN), unauthorized (or secondary) users should find unused frequency bands using spectrum sensing in order to transmit their data in those unused bands without generating harmful interference for licensed (or primary) users. To increase probability of correctly detecting unused licensed spectrum bands, a common solution is to use a cooperative spectrum sensing (CSS). However, in practice, the CSS faces the important challenge of an imperfect reporting process, in which noise and fading causing sensing results are degraded, and also generating an overhead traffic due to reporting all sensing results. In this thesis, we consider the above challenge, and propose a novel CSS strategy for CRNs with sensing multi-frequency bands simultaneously. This CSS strategy uses five techniques: (1) it uses an efficient clustering algorithm for simultaneously spectrum sensing of multiple frequency bands, without imposing a new overhead traffic and extra energy to reporting process; (2) it uses another efficient clustering algorithm to evenly distribute the total energy consumption of the CRN among clusters. As a result, it guarantees that the CRN to have maximum lifetime; (3) it employs an orthogonal distributed space-time block code (ODSTBC) to reduce the probability of errors in reporting channels. (4) In each cluster, an incremental weighting (IW) fusion rule is used to improve the accuracy of local sensing results by secondary users (SUs); (5) it selects more reliable improved sensing results for sending to the fusion center. By using this method, the overhead traffic is reduced in reporting channels while sensing performance is maintained.
Values of the key parameters of the proposed CSS (such as the number of cooperative SUs, number of clusters) affect CRN’s throughput. Therefore, first we derive closed-form expressions for the upper bound of ODSTBC bit error rate, the detection probability, and the false alarm probability of the proposed strategy. Then, we propose an optimization problem for throughput maximization of the CRN with enough protection of PUs, and we prove there is a discrete space for optimization variables. Simulation results confirm the good performance of the proposed CSS for different SNR’s in reporting channels so that it has 11% more throughput compared with that of the non-cooperative strategy, when the range of target probability of detection (to protect the PU signal) is 0.7 to 0.99.
