Face à la demande croissante de débits de données de plus en plus élevées, l’une des principales solutions proposées par la 5G est de densifier le réseau en y intégrant notamment de nouvelles « Small cells ». La réorganisation de l’architecture du réseau mobile pour s’adapter à l’intégration poussée de ces Small cells, fait naître la problématique de la connexion backhaul entre les stations de bases desservant les Small cells et le cœur de réseau. Ainsi, des liaisons backhaul de plusieurs Gb/s de données sont nécessaires pour pouvoir assurer un débit de données d’au moins 100Mb/s à l’utilisateur qui est l’un des objectifs fixés pour la 5G. Les solutions de connexion backhaul sans fils ont un avantage indiscutable face aux coûts de déploiements de fibres optiques qui sont très élevés. Pour augmenter la capacité spectrale des liaisons sans fils, l’utilisation des fréquences millimétriques au-delà de 6 GHz caractérisées par des larges bandes passantes sera prochainement discutée pour la 5G durant le World Radiocommunication Conference 2019. Parmi ces fréquences, les bandes V (57-66GHz) et E (71-76 GHz et 81-86 GHz) ont un intérêt indéniable grâce aux larges bandes passantes disponibles ainsi qu’aux conditions de licenciement peu exigeantes. Les travaux développés dans cette thèse consistent à concevoir des antennes directives à large bande passante permettant d’établir les liens backhaul point-à-point sans fils (LoS). En exploitant les technologies de fabrications à faibles coût telles que l’impression 3D et Printed Circuit Board (PCB) sur des substrats FR4, la conception de deux types d’antenne directives a été étudiée à savoir des antennes lentilles et des antennes réseaux. / In order to address the ever-increasing demand of higher data rates, adding small cells to the existing macrocells infrastructure is one of the most important milestones of the 5G roadmap. With the integration of small cells and the re-organization of the network topology, backhaul bottleneck is the main challenge to address in the near future. Facing the costs of deployments of fiber optic connections, point-to-point wireless backhaul links using millimeter wave (mmW) frequencies are gaining prominence. 5G future frequencies, to be discussed under the World Radiocommunication Conference 2019 (WRC-19) open-up the way towards mmW frequency band where large bandwidths are naturally available. The high bandwidths available at these frequencies enable several Gbps data rate backhaul links, which is un utmost necessity to respect the 100 Mbps user-experienced data rate promised by the 5G standard. Millimeter-wave frequencies in V and E-bands unlicensed/light licensed spectrum are considered as primary candidates for backhaul links. In addition to the light license regime, the high free space path loss experienced at these frequencies is rather beneficial to limit the interference between small cells links. Moreover, the high available bandwidths at V and E-bands enable to achieve multi Gb/s links without using complex modulation schemes. In this thesis, we focused our research study on developing high gain wide-band antennas usable in point-to-point backhaul links in a Line of Sight (LoS) context. Leveraging cost-efficient technologies like 3D printing and Printed Circuit Board (PCB) on FR4 substrates, we studied two high-gain antenna types: lens antennas and flat array antennas.

Conselho Nacional de Desenvolvimento Científico e Tecnológico / A espasticidade é observada na maioria dos pacientes após Acidente Vascular Encefálico (AVE), e exercem influência na presença de deficiências e incapacidades, comprometendo a função motora. As estratégias de neuroreabilitação, o biofeedback eletromiográfico (EMG), têm sido utilizado com aceitação na comunidade médica para reajustes nas habilidades sensório-motoras como retreinamento motor, redução da espasticidade e/ou treinamento de relaxamento. A intervenção ainda apresenta evidências delimitadas quanto a sua efetividade na reabilitação, principalmente, quanto a interferência sobre a atividade cortical e na redução dos sinais espásticos que oferece características negativas na execução do movimento. O objetivo dessa pesquisa foi analisar a possível interferência do treino com biofeedback eletromiográfico sobre a conscientização do controle motor no membro espástico e a caracterização da atividade das bandas de baixa frequência em diferentes regiões corticais orientada pela técnica de treinamento. Dezesseis voluntários acometidos por AVE isquêmicos foram selecionados e divididos em dois grupos (n=8). Grupo experimental (GE) submetidos ao treino com biofeedback associado a fisioterapia, e o grupo controle (GC) submetido apenas à fisioterapia convencional. Foi realizada a avaliação do grau de espasticidade pelo limiar de reflexo do estiramento tônico (LRET) e pela Escala Modificada de Ashworth (EMA) antes e três semanas após o término do treinamento com a técnica. Os sujeitos foram submetidos ao biofeedback durante 6 semanas, com 2 sessões semanalmente. O mesmo tempo para o tratamento fisioterápico foi padronizado para o GC. Os resultados mostram variação do percentual médio de melhora do grau de espasticidade, mensuradas pelo LRET, de 38,59% (dp=13,03%) no GE comparado com 18,58% (dp=11,90%) do GC. Essa variação apresentou diferença significativa (p=0,020; t=2,776; p<5%) entre os grupos (controle e experimental), e a diferença significativa do LRET antes e após do treinamento no GE (p=0,003; t=5,338; p<5%) quando comparado ao GC (p=0,015; t=3,657; p<5%). A medida semi-quantitativa da EMA antes e após o término das sessões mostraram variações apenas no GE. Com relação a atividade cortical, houve diferença de atividade das bandas (delta, teta, alfa e beta) quando a 3ª e 12ª sessão foram comparadas para cada sujeito do GE. Essa diferença foi encontrada, principalmente, em regiões frontal, central (vértex), parietal e occipital em ambos hemisférios (ipsilateral e contralateral a lesão) tanto na fase de planejamento cognitivo motor quanto na execução do movimento. Houve predominância da diferença de atividade para a banda delta, alfa e beta em diferentes sujeitos distribuída difusamente ao longo dos canais de registro de viii EEG. A atividade diferenciada das bandas foi devido ao aumento e/ou diminuição da energia espectral entre as sessões, notado apenas em alguns voluntários do GE. Observou ainda diferença de atividade em áreas motoras secundárias. As avaliações, principalmente do LRET mostram que o treino com biofeedback EMG foi efetivo na redução do grau de espasticidade. A diferença de atividade cortical das bandas de frequência entre as sessões sugere que o biofeedback modula a cognição por meio do esforço e atenção imposta pela tarefa na tentativa do movimento no membro acometido. Além de que a diferença de energia espectral entre as sessões é dependente do ajuste e complexidade da tarefa direcionado pelos sinais do biofeedback, auxiliando na aprendizagem motora. / Spasticity is observed in most patients after cerebrovascular accident (CVA), and exerts influence in the presence of disabilities, affecting motor function. For neurorehabilitation strategies, electromyographic biofeedback (EMG) has been used with acceptance in the medical community for adjustments in the sensory-motor skills as a motor retraining, reducing spasticity and/or relaxation training. The intervention still presents limited evidence regarding their effectiveness in rehabilitation, especially as the interference of cortical activity and the reduction of spastic signs that provides negative characteristics in movement execution. The aim of this study was to analyze the workout possible interference with EMG biofeedback on the motor control awareness in spastic member in the characterization of the activity of low-frequency bands in different cortical regions targeted by the training technique. Sixteen volunteers affected by ischemic stroke were selected and divided into two groups (n = 8). Experimental group (EG) underwent biofeedback training associated with physical therapy and control group (CG) only conventional physiotherapy. The assessment of the degree of spasticity by reflex threshold of the tonic stretch (TSRT) and Modified Ashworth Scale (MAS) was performed before and three weeks after the end of treatment with the technique. The subjects underwent biofeedback for 6 weeks, with two sessions weekly. The same time for physiotherapy treatment was standardized to the GC. The results show variation of the average percent improvement in the degree of spasticity measured at TRST, 38,59% (sd=13,03%) in GE compared to 18,58% (sd=11,90%) of GC. This variation showed a significant difference (p=0.020; t=2,776; p<5%) between groups (control and experimental), and the significant difference of TRST before and after training in EG (p = 0.003; t=5,338; p <5%) when compared to the CG (p=0.015; t=0,015; p<5%). The semi-quantitative measure of the MAS before and after the end of the sessions presented variations only in GE. Regarding the cortical activity, there were band activities differences when 3rd and 12th sessions were compared for each subject of GE. This difference was found primarily in the frontal, central (vertex), parietal and occipital lobe in both hemispheres (contralateral and ipsilateral to the lesion) in both the cognitive motor planning phase and in the movement execution. It was observed the predominance of activity difference for the delta band, alpha and beta in different subjects distributed diffusely over the EEG recording channels. The different activity of the bands was due to the increase and/or decrease the spectral energy between sessions, x noticed only in some GE volunteers. It was also pointed out distinct activity in secondary motor areas. Evaluations mainly from TSRT show that training with EMG biofeedback was effective in reducing the degree of spasticity. The difference in cortical activity of the frequency bands between sessions suggested that biofeedback modulates cognition through the effort and attention required by the task of movement attempt in the affected limb. Besides that, the spectral energy difference between the sessions depends on the tuning and task complexity driven by biofeedback signals, helping motor learning. / Tese (Doutorado)