Dynamic Spectrum Access (DSA) has strong potential to address the need for improved spectrum efficiency. Unfortunately, traditional DSA approaches such as simple "sense-and-avoid" fail to provide sufficient performance in many scenarios. Thus, the combination of sensing with deep reinforcement learning (DRL) has been shown to be a promising alternative to previously proposed simplistic approaches. DRL does not require the explicit estimation of transition probability matrices and prohibitively large matrix computations as compared to traditional reinforcement learning methods. Further, since many learning approaches cannot solve the resulting online Partially-Observable Markov Decision Process (POMDP), Deep Recurrent Q-Networks (DRQN) have been proposed to determine the optimal channel access policy via online learning. The fundamental goal of this dissertation is to develop DRL-based solutions to address this POMDP-DSA problem. We mainly consider three aspects in this work: (1) optimal transmission strategies, (2) combined intelligent sensing and transmission strategies, and (c) learning efficiency or online convergence speed. Four key challenges in this problem are (1) the proposed DRQN-based node does not know the other nodes' behavior patterns a priori and must to predict the future channel state based on previous observations; (2) the impact to primary user throughput during learning and even after learning must be limited; (3) resources can be wasted the sensing/observation; and (4) convergence speed must be improved without impacting performance performance. We demonstrate in this dissertation, that the proposed DRQN can learn: (1) the optimal transmission strategy in a variety of environments under partial observations; (2) a sensing strategy that provides near-optimal throughput in different environments while dramatically reducing the needed sensing resources; (3) robustness to imperfect observations; (4) a sufficiently flexible approach that can accommodate dynamic environments, multi-channel transmission and the presence of multiple agents; (5) in an accelerated fashion utilizing one of three different approaches. / Doctor of Philosophy / With the development of wireless communication, such as 5G, global mobile data traffic has experienced tremendous growth, which makes spectrum resources even more critical for future networks. However, the spectrum is an exorbitant and scarce resource. Dynamic Spectrum Access (DSA) has strong potential to address the need for improved spectrum efficiency. Unfortunately, traditional DSA approaches such as simple "sense-and-avoid" fail to provide sufficient performance in many scenarios. Thus, the combination of sensing with deep reinforcement learning (DRL) has been shown to be a promising alternative to previously proposed simplistic approaches. Compared with traditional reinforcement learning methods, DRL does not require explicit estimation of transition probability matrices and extensive matrix computations. Furthermore, since many learning methods cannot solve the resulting online partially observable Markov decision process (POMDP), a deep recurrent Q-network (DRQN) is proposed to determine the optimal channel access policy through online learning. The basic goal of this paper is to develop a DRL-based solution to this POMDP-DSA problem. This paper mainly focuses on improving performance from three directions. 1. Find the optimal (or sub-optimal) channel access strategy based on fixed partial observation mode; 2. Based on work 1, propose a more intelligent way to dynamically and efficiently find more reasonable (higher efficiency) sensing/observation policy and corresponding channel access strategy; 3. On the premise of ensuring performance, use different machine learning algorithms or structures to improve learning efficiency and avoid users waiting too long for expected performance. Through the research in these three main directions, we have found an efficient and diverse solution, namely DRQN-based technology.

Common control channels in cognitive radio (CR) ad hoc networks are spectrum resources temporarily allocated and commonly available to CR users for control message exchange. With no presumably available network infrastructure, CR users rely on cooperation to perform spectrum management functions. One the one hand, CR users need to cooperate to establish common control channels, but on the other hand, they need to have common control channels to facilitate such cooperation. This control channel problem is further complicated by primary user (PU) activities, channel impairments, and intelligent attackers. Therefore, how to reliably and securely establish control links in CR ad hoc networks is a challenging problem. In this work, a framework for control channel design and analysis is proposed to address control channel reliability and security challenges for seamless communication and spectral efficiency in CR ad hoc networks. The framework tackles the problem from three perspectives: (i) responsiveness to PU activities: an efficient recovery control channel method is devised to efficiently establish control links and extend control channel coverage upon PU's return while mitigating the interference with PUs, (ii) robustness to channel impairments: a reinforcement learning-based cooperative sensing method is introduced to improve cooperative gain and mitigate cooperation overhead, and (iii) resilience to jamming attacks: a jamming-resilient control channel method is developed to combat jamming under the impacts of PU activities and spectrum sensing errors by leveraging intrusion defense strategies. This research is particularly attractive to emergency relief, public safety, military, and commercial applications where CR users are highly likely to operate in spectrum-scarce or hostile environment.

L’allocation des ressources spectrales à des services de communications sans fil, sans cesse plus nombreux et plus gourmands, a récemment mené la communauté radio à vouloir remettre en question la stratégie de répartition des bandes de fréquences imposée depuis plus d’un siècle. En effet une étude rendue publique en 2002 par la commission fédérale des communications aux Etats-Unis (Federal Communications Commission - FCC) mit en évidence une pénurie des ressources spectrales dans une large bande de fréquences comprise entre quelques mégahertz à plusieurs gigahertz. Cependant, cette même étude expliqua cette pénurie par une allocation statique des ressources aux différents services demandeurs plutôt que par une saturation des bandes de fréquences. Cette explication fut par la suite corroborée par de nombreuses mesures d’occupation spectrale, réalisées dans plusieurs pays, qui montrèrent une forte sous-utilisation des bandes de fréquences en fonction du temps et de l’espace, représentant par conséquent autant d’opportunité spectrale inexploitée. Ces constations donnèrent naissance à un domaine en plein effervescence connu sous le nom d’Accès Opportuniste au Spectre (Opportunistic Spectrum Access). Nos travaux suggèrent l’étude de mécanismes d’apprentissage pour la radio intelligente (Cognitive Radio) dans le cadre de l’Accès Opportuniste au Spectre (AOS) afin de permettre à des équipements radio d’exploiter ces opportunités de manière autonome. Pour cela, nous montrons que les problématiques d’AOS peuvent être fidèlement représentées par des modèles d’apprentissage par renforcement. Ainsi, l’équipement radio est modélisé par un agent intelligent capable d’interagir avec son environnement afin d’en collecter des informations. Ces dernières servent à reconnaître, au fur et à mesure des expériences, les meilleurs choix (bandes de fréquences, configurations, etc.) qui s’offrent au système de communication. Nous nous intéressons au modèle particulier des bandits manchots (Multi-Armed Bandit appliqué à l’AOS). Nous discutons, lors d’une phase préliminaire, différentes solutions empruntées au domaine de l’apprentissage machine (Machine Learning). Ensuite, nous élargissons ces résultats à des cadres adaptés à la radio intelligente. Notamment, nous évaluons les performances de ces algorithmes dans le cas de réseaux d’équipements qui collaborent en prenant en compte, dans le modèle suggéré, les erreurs d’observations. On montre de plus que ces algorithmes n’ont pas besoin de connaître la fréquence des erreurs d’observation afin de converger. La vitesse de convergence dépend néanmoins de ces fréquences. Dans un second temps nous concevons un nouvel algorithme d’apprentissage destiné à répondre à des problèmes d’exploitation des ressources spectrales dans des conditions dites de fading. Tous ces travaux présupposent néanmoins la capacité de l’équipement intelligent à détecter efficacement l’activité d’autres utilisateurs sur la bande (utilisateurs prioritaires dits utilisateurs primaires). La principale difficulté réside dans le fait que l’équipement intelligent ne suppose aucune connaissance a priori sur son environnement (niveau du bruit notamment) ou sur les utilisateurs primaires. Afin de lever le doute sur l’efficacité de l’approche suggérée, nous analysons l’impact de ces incertitudes sur le détecteur d’énergie. Ce dernier prend donc le rôle d’observateur et envoie ses observations aux algorithmes d’apprentissage. Nous montrons ainsi qu’il est possible de quantifier les performances de ce détecteur dans des conditions d’incertitude sur le niveau du bruit ce qui le rend utilisable dans le contexte de la radio intelligente. Par conséquent, les algorithmes d’apprentissage utilisés pourront exploiter les résultats du détecteur malgré l’incertitude inhérente liée à l’environnement considéré et aux hypothèses (sévères) d’incertitude liées au problème analysé. / During the last century, most of the meaningful frequency bands were licensed to emerging wireless applications. Because of the static model of frequency allocation, the growing number of spectrum demanding services led to a spectrum scarcity. However, recently, series of measurements on the spectrum utilization showed that the different frequency bands were underutilized (sometimes even unoccupied) and thus that the scarcity of the spectrum resource is virtual and only due to the static allocation of the different bands to specific wireless services. Moreover, the underutilization of the spectrum resource varies on different scales in time and space offering many opportunities to an unlicensed user or network to access the spectrum. Cognitive Radio (CR) and Opportunistic Spectrum Access (OSA) were introduced as possible solutions to alleviate the spectrum scarcity issue.In this dissertation, we aim at enabling CR equipments to exploit autonomously communication opportunities found in their vicinity. For that purpose, we suggest decision making mechanisms designed and/or adapted to answer CR related problems in general, and more specifically, OSA related scenarios. Thus, we argue that OSA scenarios can be modeled as Multi-Armed Bandit (MAB) problems. As a matter of fact, within OSA contexts, CR equipments are assumed to have no prior knowledge on their environment. Acquiring the necessary information relies on a sequential interaction between the CR equipment and its environment. Finally, the CR equipment is modeled as a cognitive agent whose purpose is to learn while providing an improving service to its user. Thus, firstly we analyze the performance of UCB1 algorithm when dealing with OSA problems with imperfect sensing. More specifically, we show that UCB1 can efficiently cope with sensing errors. We prove its convergence to the optimal channel and quantify its loss of performance compared to the case with perfect sensing. Secondly, we combine UCB1 algorithm with collaborative and coordination mechanism to model a secondary network (i.e. several SUs). We show that within this complex scenario, a coordinated learning mechanism can lead to efficient secondary networks. These scenarios assume that a SU can efficiently detect incumbent users’ activity while having no prior knowledge on their characteristics. Usually, energy detection is suggested as a possible approach to handle such task. Unfortunately, energy detection in known to perform poorly when dealing with uncertainty. Consequently, we ventured in this Ph.D. to revisit the problem of energy detection limits under uncertainty. We present new results on its performances as well as its limits when the noise level is uncertain and the uncertainty is modeled by a log-normal distribution (as suggested by Alexander Sonnenschein and Philip M. Fishman in 1992). Within OSA contexts, we address a final problem where a sensor aims at quantifying the quality of a channel in fading environments. In such contexts, UCB1 algorithms seem to fail. Consequently, we designed a new algorithm called Multiplicative UCB (UCB) and prove its convergence. Moreover, we prove that MUCB algorithms are order optimal (i.e., the order of their learning rate is optimal). This last work provides a contribution that goes beyond CR and OSA. As a matter of fact, MUCB algorithms are introduced and solved within a general MAB framework.

Given the ever increasing reliance of today’s society on ubiquitous wireless access, the paradigm of dynamic spectrum access (DSA) as been proposed and implemented for utilizing the limited wireless spectrum more efficiently. Orthogonal frequency division multiplexing (OFDM) is growing in popularity for adoption into wireless services employing DSA frame- work, due to its high bandwidth efficiency and resiliency to multipath fading. While these advantages have been proven for many wireless applications, including LTE-Advanced and numerous IEEE wireless standards, one potential drawback of OFDM or its non-contiguous variant, NC-OFDM, is that it exhibits high peak-to-average power ratios (PAPR), which can induce in-band and out-of-band (OOB) distortions when the peaks of the waveform enter the compression region of the transmitter power amplifier (PA). Such OOB emissions can interfere with existing neighboring transmissions, and thereby severely deteriorate the reliability of the DSA network. A performance-enhancing digital pre-distortion (DPD) technique compensating for PA and in-phase/quadrature (I/Q) modulator distortions is proposed in this dissertation. Al- though substantial research efforts into designing DPD schemes have already been presented in the open literature, there still exists numerous opportunities to further improve upon the performance of OOB suppression for NC-OFDM transmission in the presence of RF front-end impairments. A set of orthogonal polynomial basis functions is proposed in this dissertation together with a simplified joint DPD structure. A performance analysis is presented to show that the OOB emissions is reduced to approximately 50 dBc with proposed algorithms employed during NC-OFDM transmission. Furthermore, a novel and intuitive DPD solution that can minimize the power regrowth at any pre-specified frequency in the spurious domain is proposed in this dissertation. Conventional DPD methods have been proven to be able to effectively reduce the OOB emissions that fall on top of adjacent channels. However more spectral emissions in more distant frequency ranges are generated by employing such DPD solutions, which are potentially in violation of the spurious emission limit. At the same time, the emissions in adjacent channel must be kept under the OOB limit. To the best of the author’s knowledge, there has not been extensive research conducted on this topic. Mathematical derivation procedures of the proposed algorithm are provided for both memoryless nonlinear model and memory-based nonlinear model. Simulation results show that the proposed method is able to provide a good balance of OOB emissions and emissions in the far out spurious domain, by reducing the spurious emissions by 4-5 dB while maintaining the adjacent channel leakage ratio (ACLR) improvement by at least 10 dB, comparing to the PA output spectrum without any DPD.

Cognitive radio is a recent novel approach for the realization of intelligent and sophisticated wireless systems. Although the research and development on cognitive radio is still in the stage of infancy, there are significant interests and efforts towards realization of cognitive radio. Cognitive radio systems are envisioned to support context awareness and related systems. The context can be spectrum, environment, location, waveform, power and other radio resources. Significant amount of the studies related to cognitive radio in the literature focuses on the spectrum awareness since it is one of the most crucial features of cognitive radio systems. However, the rest of the features of cognitive radio such as location and environment awareness have not been investigated thoroughly. For instance, location aware systems are widespread and the demand for more advanced ones are growing. Therefore, the main objective of this dissertation is to develop an underlying location awareness architecture for cognitive radio systems, which is described as location awareness engine, in order to support goal driven and autonomous location aware systems. A cognitive radio conceptual model with location awareness engine and cycle is developed by inspiring from the location awareness features of human being and bat echolocation systems. Additionally, the functionalities of the engine are identified and presented. Upon providing the functionalities of location awareness engine, the focus is given to the development of cognitive positioning systems. Furthermore, range accuracy adaptation, which is a cognitive behavior of bats, is developed for cognitive positioning systems. In what follows, two main approaches are investigated in order to improve the performance of range accuracy adaptation method. The first approach is based on idea of improving the spectrum availability through hybrid underlay and overlay dynamic spectrum access method. On the other hand, the second approach emphasizes on spectrum utilization, where we study performance of range accuracy adaptation from both theoretical and practical perspectives considering whole spectrum utilization approach. Furthermore, we introduced a new spectrum utilization technique that is referred as dispersed spectrum utilization. The performance analysis of dispersed spectrum utilization approach is studied considering time delay estimation problem in cognitive positioning systems. Afterward, the performance of whole and dispersed spectrum utilization approaches are compared in the context of cognitive positioning systems. Finally, some representative advanced location aware systems for cognitive radio networks are presented in order to demonstrate some potential applications of the proposed location awareness engine in cognitive radio systems.

Cognitive Radio (CR) is one possible option for mitigating the inefficient wireless spectrum distribution that occurs as a result of fixed spectrum allocation. The use of Dynamic Spectrum Access capabilities will potentially enable secondary users to utilize available and unoccupied frequency slots (channels) whenever the licensed users for those channels are absent. In Cognitive Radio Networks (CRNs), whenever users access the spectrum in an opportunistic manner, control messaging is a crucial issue to ensure that secondary users, i.e. Cognitive Radio Users (CRUs), do not interfere with the licensed users, i.e. Primary Users. In CRNs, where not all CRUs share the same set of channels, i.e. CRUs with Heterogeneous Frequency Devices (HFD), a set of channels must be chosen with care to allow all CRUs in the network to be able to transmit and receive control information. The thesis considers how Control Messaging Schemes (CMSs) can be used within CRNs and proposes a novel CMS for a CRN supporting HFDs. The thesis starts by classifying the CMSs; generating a new taxonomy and identifying the main characteristics for an efficient CRN with HFD. Then, different mathematical approaches for choosing the set of channels used for control information are presented. Next, a CMS for a CRN with HFDs model based upon the aforementioned characteristics and calculating the minimum number of channels for transmitting control information is proposed. Finally the thesis concludes with a number of CMS being presented and evaluated in terms of their impact upon transmission efficiency.

As more and more wireless applications/services emerge in the market, the already heavily crowded radio spectrum becomes much scarcer. Meanwhile, however,as it is reported in the recent literature, there is a large amount of radio spectrum that is under-utilized. This motivates the concept of cognitive radio wireless networks that allow the unlicensed secondary-users (SUs) to dynamically use the vacant radio spectrum which is not being used by the licensed primary-users (PUs). In this dissertation, we investigate protocol design for both the synchronous and asynchronous cognitive radio networks with emphasis on the medium access control (MAC) layer. We propose various spectrum sharing schemes, opportunistic packet scheduling schemes, and spectrum sensing schemes in the MAC and physical (PHY) layers for different types of cognitive radio networks, allowing the SUs to opportunistically utilize the licensed spectrum while confining the level of interference to the range the PUs can tolerate. First, we propose the cross-layer based multi-channel MAC protocol, which integrates the cooperative spectrum sensing at PHY layer and the interweave-based spectrum access at MAC layer, for the synchronous cognitive radio networks. Second, we propose the channel-hopping based single-transceiver MAC protocol for the hardware-constrained synchronous cognitive radio networks, under which the SUs can identify and exploit the vacant channels by dynamically switching across the licensed channels with their distinct channel-hopping sequences. Third, we propose the opportunistic multi-channel MAC protocol with the two-threshold sequential spectrum sensing algorithm for asynchronous cognitive radio networks. Fourth, by combining the interweave and underlay spectrum sharing modes, we propose the adaptive spectrum sharing scheme for code division multiple access (CDMA) based cognitive MAC in the uplink communications over the asynchronous cognitive radio networks, where the PUs may have different types of channel usage patterns. Finally, we develop a packet scheduling scheme for the PU MAC protocol in the context of time division multiple access (TDMA)-based cognitive radio wireless networks, which is designed to operate friendly towards the SUs in terms of the vacant-channel probability. We also develop various analytical models, including the Markov chain models, M=GY =1 queuing models, cross-layer optimization models, etc., to rigorously analyze the performance of our proposed MAC protocols in terms of aggregate throughput, access delay, and packet drop rate for both the saturation network case and non-saturation network case. In addition, we conducted extensive simulations to validate our analytical models and evaluate our proposed MAC protocols/schemes. Both the numerical and simulation results show that our proposed MAC protocols/schemes can significantly improve the spectrum utilization efficiency of wireless networks.

Les systèmes de communications sans fil ne cessent de se multiplier pour devenir incontournables de nos jours. Cette croissance cause une augmentation de la demande des ressources spectrales, qui sont devenues de plus en plus rares. Afin de résoudre ce problème de pénurie de fréquences, Joseph Mitola III, en 2000, a introduit l'idée de l'allocation dynamique du spectre. Il définit ainsi le terme " Cognitive Radio " (Radio Intelligente), qui est largement pressenti pour être le prochain Big Bang dans les futures communications sans fil [1]. Dans le cadre de ce travail on s'intéresse à la problématique du spectrum sensing qui est la détection de présence des Utilisateurs Primaires dans un spectre sous licence, dans le contexte de la radio intelligente. L'objectif de ce travail est de proposer des méthodes de détection efficaces à faible complexité et/ou à faible temps d'observation et ceci en utilisant le minimum d'information a priori sur le signal à détecter. Dans la première partie on traite le problème de détection d'un signal aléatoire dans le bruit. Deux grandes méthodes de détection sont utilisées : la détection d'énergie ou radiomètre et la détection cyclostationnaire. Dans notre contexte, ces méthodes sont plus complémentaires que concurrentes. Nous proposons une architecture hybride de détection des bandes libres, qui combine la simplicité du radiomètre et la robustesse des détecteurs cyclostationnaires. Deux méthodes de détection sont proposées qui se basent sur cette même architecture. Grâce au caractère adaptatif de l'architecture, la détection évolue au cours du temps pour tendre vers la complexité du détecteur d'énergie avec des performances proches du détecteur cyclostationnaire ou du radiomètre selon la méthode utilisée et l'environnement de travail. Dans un second temps on exploite la propriété parcimonieuse de la Fonction d'Autocorrelation Cyclique (FAC) pour proposer un nouvel estimateur aveugle qui se base sur le compressed sensing afin d'estimer le Vecteur d'Autocorrelation Cyclique (VAC), qui est un vecteur particulier de la Fonction d'Autocorrelation Cyclique pour un délai fixe. On montre par simulation que ce nouvel estimateur donne de meilleures performances que celles obtenues avec l'estimateur classique, qui est non aveugle et ceci dans les mêmes conditions et en utilisant le même nombre d'échantillons. On utilise l'estimateur proposé, pour proposer deux détecteurs aveugles utilisant moins d'échantillons que nécessite le détecteur temporel de second ordre de [2] qui se base sur l'estimateur classique de la FAC. Le premier détecteur exploite uniquement la propriété de parcimonie du VAC tandis que le second détecteur exploite en plus de la parcimonie la propriété de symétrie du VAC, lui permettant ainsi d'obtenir de meilleures performances. Ces deux détecteurs outre qu'ils sont aveugles sont plus performants que le détecteur non aveugle de [2] dans le cas d'un faible nombre d'échantillons.

Le nombre croissant d'applications des Réseaux de Capteurs Sans Fils (RCSFs) a conduit les industriels à concevoir ces réseaux avec une couche Physique (PHY) suivant le standard IEEE 802.15.4. Actuellement, cette couche est implémentée en matériel souffrant d'un manque de flexibilité du changement des paramètres radio, telles que bandes de fréquences et modulations. Ce problème est accentué par la rareté du spectre radio fréquences. La Radio Logiciel (RL) est une nouvelle solution pour reconfigurer plus facilement ces paramètres. A partir d'une RL, il est possible de développer une radio cognitive permettant une écoute de spectre et un Accès Dynamique au Spectre (ADS). Ces deux possibilités sont utiles pour surmonter le problème de la rareté du spectre. Cette thèse propose une nouvelle solution Radio logicielle pour un RCSF basé sur le standard IEEE 802.15.4. Notre objectif est de caractériser une plate-forme RL qui implémente à la fois deux couches PHY standardisées et une radio cognitive pour des RCSFs. Dans cette thèse, nous avons réalisé des implémentations RL en utilisant une plateforme composée de la solution Universal Software Peripheral Radio (USRP) d'Ettus Research et de GNU Radio. Nous avons choisi cette plateforme particulière puisqu'elle est parmi les outils les plus performants et les plus pratiques d'après notre état de l'art. Une étude minutieuse a été effectuée pour analyser l'architecture logicielle de la GNU Radio avant son utilisation. Des USRPs avec leurs cartes filles ont été aussi analysés à travers des mesures expérimentales radio fréquences. L'analyse de cette plate-forme a apporté une description détaillée de son architecture et de ses performances. Nous avons prouvé que les performances mesurées sont plus faibles que ceux attendus pour certaines cartes filles d'USRP. Malgré ces résultats, certaines cartes ont de nombreuses caractéristiques intéressantes, comme de grandes bandes de fréquences couvertes et une puissance de sortie linéaire. Un modèle empirique a été introduit pour caractériser avec précision la puissance de sortie moyenne d'une carte fille particulière. Nous avons ensuite implémenté une nouvelle couche PHY standardisée pour la bande de fréquence 868/915 MHz basée sur le standard 802.15.4. Un processus de rétro-ingénierie d'une autre implémentation développée pour la bande 2.4GHz a été effectué. Ces deux couches ont été décrites par des chaines de communications ou des graphes de flux. Nous avons finalement proposé une nouvelle radio cognitive par une reconfiguration de ces graphes de flux dans les deux bandes de fréquences correspondantes. La particularité de notre radio cognitive est de reconfigurer les graphes de flux en fonction de la fréquence sélectionnée. Cette sélection est effectuée par un ADS et une écoute de spectre basé sur une détection d'énergie, validés tous les deux au travers des réelles communications sans fil. Nous avons introduit un algorithme à base de messages afin de reconfigurer les graphes de flux et de synchroniser la sélection sur une fréquence porteuse. Les deux couches physiques en RL pour les bandes 2.4 GHz et 868/915 MHz ont été testées et sont fonctionnelles. La première a été testée en échangeant des paquets de données avec des nœuds capteurs réels. La deuxième a été expérimentée par l'échange de paquets, à travers une communication entre deux radios logicielles USRP/GNU Radio. Nous avons réussi à mesurer deux paramètres réels d'une communication sans fil : le taux d'erreur binaire et le taux de succès des paquets. Les couches PHY résultantes ont servi à la réalisation et à l'expérimentation d'un ADS de notre radio cognitive. Un ADS a amélioré significativement le taux de succès de paquets par rapport à celui obtenu avec un accès statique dans un environnement indoor. Les résultats de cette thèse conduisent à expérimenter une radio cognitive avec une RL non seulement pour un RCSF, mais pour d'autres réseaux sans fil et standards radio / The Increasing number of Wireless Sensor Networks (WSNs) applications has led industries to design the physical layer (PHY) of these networks following the IEEE 802.15.4 standard. The traditional design of that layer is on hardware suffering from a lack of flexibility of radio parameters, such as changing both frequency bands and modulations. This problem is emphasized by the scarcity of the radio-frequency spectrum. Software Defined Radio (SDR) is an attracting solution to easily reconfigure radio parameters. In addition to SDR, a cognitive radio concept can be proposed by spectrum sensing and Dynamic Spectrum Access (DSA) both to overcome the spectrum scarcity problem. This thesis proposes a new SDR solution for WSNs based on the IEEE 802.15.4 standard. Our aim is to characterize an SDR platform that implements two standardized PHY layers and cognitive radio for WSNs. In this thesis, we carried out SDR implementations using a GNU Radio and Universal Software Peripheral Radio (USRP) platform. We chose this particular platform because it is one of the most practical and well-performed ones. A thorough study was performed to analyze GNU Radio software architecture before its usage. USRPs and their daughter boards were also analyzed through experimental radio-frequency measurements. The analysis of the GNU Radio USRP platform brought a detailed description of its architecture and performances as well as the way to implement an SDR. This description particularly assists researchers to quickly develop efficient SDR receivers and transmitters. We show through our experiments that the measured performances of daughter boards mounted on a USRP are lower than expected ones. Despite these results, some daughter boards have many interesting features such as large covered frequency bands and with a linear output power. An empirical model was introduced to accurately characterize the average output power of a particular daughter board. Then, we implemented a new possible standardized PHY layer for the 868/915 MHz frequency band. A reverse engineering process of another implementation was performed for the 2450 MHz frequency band. These two PHY layers were described by communication chains or flow graphs. We suggested a new Cognitive Radio by a reconfiguration of these flow graphs within the corresponding frequency bands. The particularity of our cognitive radio is to reconfigure flow graphs in function to the selected frequency. This selection is performed by DSA and spectrum sensing based on energy detection both through real wireless communications. We introduced a message based algorithm in order to reconfigure the flow graphs and to synchronize the selection of a carrier frequency. Our two implemented PHY layers for the 2450 MHz and the 868/915 frequency bands were found functional. The first one was tested by exchanging data packets with real sensor nodes. The second was also experienced by a packet exchange, but via GNURadio/USRP communications. Both tests were carried out through real communications. We were also able to measure two wireless communication parameters: Bit Error Rate (BER) and the Packet Success Rate (PSR). The result of functional PHY layers was beneficial for realization and experiments of our cognitive radio. We found that our DSA significantly improves the packet success rate compared to that obtained with static spectrum access in an indoor environment. The results of this thesis lead to experiment a cognitive radio with an SDR not only for a WSN, but for other wireless networks and radio standards

Les années récentes ont vu l’explosion du trafic sur les réseaux mobiles depuis l’apparition de nouveaux terminaux (smartphones, tablettes) et des usages qu’ils permettent, en particulier les données multimédia, le trafic voix restant sensiblement constant. Une conséquence est le besoin de plus de spectre, ou la nécessité de mieux utiliser le spectre déjà alloué. Comme il n’y a pas de coordination entre les utilisateurs secondaire(s) et primaire, avant toute transmission les premiers doivent mettre en œuvre des traitements pour détecter les périodes dans lesquelles l’utilisateur primaire transmet, ce qui est le scénario considéré dans cette thèse. Nous considérons donc une autre approche, reposant sur l’utilisation de codes correcteurs d’effacements en mode paquet. La dernière partie de la thèse aborde un scénario dans lequel il n’y a plus d’utilisateur primaire, tous les utilisateurs ayant le même droit à transmettre dans le canal. Nous décrivons une modification de la couche MAC du 802.11 consistant à réduire les différents temps consacrés à attendre (SIFS, DIFS, backoff, . . .) afin d’accéder plus souvent au canal, au prix de quelques collisions supplémentaires qu’il est possible de récupérer en mettant en œuvre des codes correcteurs d’effacements. / The emergence of new devices especially the smartphones and tablets having a lot of new applications have rocketed the wireless traffic in recent years and this is the cause of main surge in the demand of radio spectrum. There is a need of either more spectrum or to use existing spectrum more efficiently due to dramatic increase in the demand of limited spectrum. Among the new dynamic access schemes designed to use the spectrum more efficiently opportunistic spectrum access (OSA) is currently addressed when one or more secondary users (SU) are allowed to access the channel when the PU is not transmitting. The erasure correcting codes are therefore envisioned to recover the lost data due to sensing impairments. We define the parameter efficiency of SU and optimize it in-terms of spectrum utilization keeping into account sensing impairments, code parameters and the activity of PU. Finally, the spectrum access for multiple secondary users is addressed when there is no primary and each user has equal right to access the channel. The interesting scenarios are Cognitive radio networks and WiFi where 802.11 protocol gives the specification for MAC layer. The throughput curvesachieved by retransmission and using various erasure correcting codes are compared. This modification in MAC layer will reduce the long waiting time to access the channel, as the number of users are increased.