The demand of wireless spectrum is increasing very fast as the field of telecommunication is advancing rapidly. The spectrum was underutilized because of fixed spectrum assignment policy and this valuable spectrum can be utilized efficiently by cognitive radio technology. In this thesis we have studied spectrum selection problems in cognitive radio network. Channel sharing and channel contention problems arise when multiple secondary users tend to select same channel. The thesis work is focused on spectrum selection issue with the aim to minimize the overall system time and to solve the problem of channel contention and channel sharing. The overall system time of secondary connection is an important performance measure to provide quality of service for secondary users in cognitive radio network. We studied two spectrum selection schemes that considerably reduce the overall system time and resolve the problems of channel sharing and channel contention. An analytical model associated with Preemptive Resume Priority (PRP) M/G/1 queuing model has been provided to evaluate the studied spectrum selection scheme. This model also analyzes the effect of multiple handoffs due to arrival of primary users. According to this scheme, the traffic load is distributed among multiple channels to balance the traffic load. Secondary users select the operating channels based on the spectrum selection algorithm. They can intelligently adopt better channel selection scheme by considering traffic statistics and overall transmission time. All simulation scenarios are developed in MATLAB. Based on our result we can conclude that both channel selection schemes considerably reduce the overall transmission time of secondary users in cognitive radio network. The overall transmission time increase with the rise of arrival rate of secondary user. The probability based channel selection scheme perform better with lower arrival rate and sensing based channel selection scheme perform better with higher arrival rate of secondary users. These channel selection schemes help distribute the traffic load of secondary users evenly among multiple channels. Hence, increase the channel utilization and resolve the channel contention problem.

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2014-11-28T12:18:45Z No. of bitstreams: 2 Dissertação - Camila Soares Barbosa - 2013.pdf: 840210 bytes, checksum: a7c84142e9c6b7a669f16d9057771acf (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-11-28T13:14:26Z (GMT) No. of bitstreams: 2 Dissertação - Camila Soares Barbosa - 2013.pdf: 840210 bytes, checksum: a7c84142e9c6b7a669f16d9057771acf (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-11-28T13:14:26Z (GMT). No. of bitstreams: 2 Dissertação - Camila Soares Barbosa - 2013.pdf: 840210 bytes, checksum: a7c84142e9c6b7a669f16d9057771acf (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2013-11-22 / One of the main challenges in Dynamic Spectrum Selection for Cognitive Radios is the choice of the frequency range for each transmission. This choice should minimize interference with legacy devices and maximize the discovering opportunities or white spaces. There are several solutions to this issue, and Reinforcement Learning algorithms are the most successful. Among them stands out the Q-Learning whose weak point is the parameterization, since adjustments are needed in order to reach successfully the proposed objective. In that sense, this work proposes an algorithm based on evolutionary strategy and presents the main characteristics adaptability to the environment and fewer parameters. Through simulation, the performance of the Q-Learning and the proposal of this work were compared in different scenarios. The results allowed to evaluate the spectral efficiency and the adaptability to the environment. The proposal of this work shows promising results in most scenarios. / Um dos principais desafios da Seleção Dinâmica de Espectro em Rádios Cognitivos é a escolha da faixa de frequência para cada transmissão. Essa escolha deve minimizar a interferência em dispositivos legados e maximizar a descoberta das oportunidades ou espaços em branco. Há várias soluções para essa questão, sendo que algoritmos de Aprendizado por Reforço são as mais bem sucedidas. Entre eles destaca-se o Q-Learning, cujo ponto fraco é a parametrização, uma vez que ajustes são necessários para que se alcance, com sucesso, o objetivo proposto. Nesse sentido, este trabalho propõe um algoritmo baseado em Estratégia Evolutiva e apresenta como características principais a adaptabilidade ao ambiente e a menor quantidade de parâmetros. Através de simulação, o desempenho do Q-Learning e da proposta deste trabalho foram comparados em diversos cenários. Os resultados obtidos permitiram avaliar a eficiência espectral e a adaptabilidade ao ambiente. A proposição deste trabalho apresentou resultados promissores na maioria dos cenários.