Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2018-10-25T12:55:35Z No. of bitstreams: 2 Dissertação - Henrique Valle de Lima - 2018.pdf: 6140413 bytes, checksum: 4bb41779722574af9e9b07103ad3ba20 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-10-25T13:11:28Z (GMT) No. of bitstreams: 2 Dissertação - Henrique Valle de Lima - 2018.pdf: 6140413 bytes, checksum: 4bb41779722574af9e9b07103ad3ba20 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-10-25T13:11:28Z (GMT). No. of bitstreams: 2 Dissertação - Henrique Valle de Lima - 2018.pdf: 6140413 bytes, checksum: 4bb41779722574af9e9b07103ad3ba20 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-09-28 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / LTE (Long Term Evolution) in unlicensed band (LTE-U) has emerged as a promising solution to the problem of the huge growth in mobile data traffic. It expands the benefits of LTE with bands of the unlicensed 5 GHz spectrum, mainly used in IEEE 802.11. But uncertainties as to the availability of these bands make the adoption of LTE-U a great challenge. In this dissertation, we propose a mixed linear programming approach for allocating resources in order to expand service by LTE-U. Subsequently, we propose a stochastic programming approach, taking into account the randomness of the unlicensed band, allocating resources under a certain control in the risk of conflict with IEEE-802.11. We used three models of the literature for the demand for IEEE 802.11 in our tests. The results highlight the importance of prior knowledge about the distribution of this demand. / A LTE (Long Term Evolution) em banda não-licenciada (LTE-U) surgiu como uma solução promissora para o problema do enorme crescimento no tráfego de dados móveis. Ela amplia os benefícios da LTE com bandas do espectro não-licenciado de 5 GHz, usado principalmente em IEEE 802.11. Mas incertezas quanto à disponibilidade dessas bandas tornam a adoção da LTE-U um grande desafio. Nesta dissertação, propomos uma abordagem de programação linear mista, para alocação de recursos a fim de expandir o atendimento por LTE-U. Posteriormente, propomos uma abordagem de programação estocástica, levando em consideração a aleatoriedade da banda não-licenciada, alocando recursos sob um certo controle no risco de conflito com o IEEE 802.11. Usamos três modelos da literatura para a demanda por IEEE 802.11 em nossos testes. Os resultados destacam a importância do conhecimento prévio sobre a distribuição dessa demanda.

Long-Term Evolution (LTE) is the fourth generation (4G) wireless communications standard and its evolution is paving the path for the fifth generation (5G) technology. LTE is also considered for supporting public safety networks, Machine-to-Machine (M2M) communications, and many other applications. Hence, it is critical to ensure that the LTE system performs effectively even in harsh signaling environments. Unfortunately, LTE is vulnerable to intentional interference at the physical layer. We define the term LTE control channel spoofing, which refers to the case when an adversary sets a fake LTE-like base station (evolved NodeB or eNodeB) that transmits a partial or full LTE downlink frame to deceive LTE devices and hinder them from attaching to a real cell. Based on analyzing the initial cell selection process in the LTE specifications, we identify three different level of LTE control channel spoofing. We have built a testbed to demonstrate the feasibility of such an attack. The experimental results show that LTE control channel spoofing can cause permanent denial of service for LTE devices during the cell selection process. We propose effective mitigation techniques to enhance the immunity of LTE systems against all the three forms of LTE control channel spoofing, and ensure that it is secure and available when and where needed. Moreover, the commercial success of LTE and the resulting growth in mobile data demand have motivated cellular network operators to strive for new innovations. LTE-Unlicensed has been recently proposed to allow cellular network operators to offload some of their data traffic by accessing the unlicensed 5 GHz frequency band. There are three variants of LTE-Unlicensed that have been proposed in the industry. These variants differ in their operational features, but they enhance the capacity of LTE and represent a big milestone in its evolution toward 5G. However, LTE-Unlicensed faces several challenges when operating in the 5 GHz bands, as this spectrum is mainly occupied by Wi-Fi and by various radar systems. Therefore, we analyze the algorithms proposed in the industry for the LTE-Unlicensed and Wi-Fi coexistence, and we develop a new spectrum sharing technique for the coexistence between LTE-Unlicensed and radar systems. In order to analyze LTE-Unlicensed and Wi-Fi coexistence, we first explain the technical details of each of the three variants of LTE-Unlicensed, and we provide a comparative analysis of them in terms of their operational features. Then we develop an unbiased and objective evaluation of their proposed coexistence mechanisms with Wi-Fi systems, and numerically compare their performance. In order to emphasize the need for developing a new spectrum sharing technique for the coexistence between LTE-Unlicensed and radar systems, we first present the different regulatory requirements for the 5 GHz unlicensed bands in several world regions, and we perform a comprehensive survey on the different radar types within the 5 GHz sub-bands. Then we develop a novel spectrum sharing technique based on chance-constrained stochastic optimization to allow the LTE-Unlicensed eNodeB to share the spectrum efficiently with a radar system. The optimization problem is formulated to guarantee the minimum performance criteria for the radar operation, and at the same time allows the LTE-Unlicensed eNodeB to control its transmit power to maximize the performance for the serving LTE-Unlicensed device. A mathematical model is used to transform the stochastic optimization problem into a deterministic one, and an exhaustive search is used to solve the resulting optimization problem. Due to the power control mechanism resulting from the proposed algorithm, numerical results show a significant reduction in the protection distance required between the radar and the LTE-Unlicensed network for the two to coexist, as the proposed algorithm can allow the two systems to operate effectively with a protection distance of only 3.95% of the one imposed by the regulations.