In today's competitive market, service providers seek to increase their revenues while keeping capital and operational expenditures down. This entails the provision of new revenue-generating multimedia services and guaranteed bandwidth for business- and mission-critical applications. It also means a requirement to consolidate disparate networks into a single infrastructure, eliminating the need to maintain several physical networks. With various networking technologies around, integrating them into one seamless infrastructure is a big challenge. The main theme of this thesis is to investigate this issue from the viewpoint of a service provider. Designing such an infrastructure would require an understanding of the properties of the traffic generated by these services in order to determine the network resources required. The issues of traffic modelling investigated leads to this thesis developing a new multi-layer modelling approach. This work provides insight into future cross-layer optimisation approach. These emerging services also require varying performance guarantees from the network. Hence the network must be QoS-aware. For IP-based networks, IntServ and DiffServ are the two QoS architectures proposed. This thesis discusses these approaches and proposes an integrated architecture to achieve a scalable end-to-end connectivity. On the broadband front, ATM is widely deployed in the core/backbone of high-speed networks for its support for QoS for real-time services. MPLS was initially developed as a new transport platform promising the robustness of IP routing at connection-oriented switching speeds. It has met this goal and currently is finding new found importance in IP traffic engineering, an important tool in achieving the QoS, performance guarantees and operational efficiency. This thesis proposes the addition of both MPLS and ATM to the end-to-end architectural framework for the new generation IP QoS. It also looks at how IPv6 can play an important role towards the evolution of this architecture.

With convergence of the Internet and mobile telecommunications, mobile networks are taking more and more IP-based protocols onboard. Mobile networks are expected to evolve and become IP-based packet data networks with various fuctionalities supported by IP-based protocols. This thesis investigates IP protocols for mobility management an all-IP mobile network. The design principles of mobility management in packet data network are first explored by studying GPRS. It is realised that GPRS is only an emulated IP network and lacks of the flexibility of IP networks. Packet scheduling techniques for enhancing mobility management performance and its implementation in GPRS are also studied. Evolution of GPRS to an all-IP based network, in which IP protocols can be used for mobility management, is subsequently proposed. Mobility management protocols in IP networks can be segregated into inter-domain (or macro) mobility, intra-domain (or micro) mobility and mobile ad hoc networking. Mobile IP is a proposal by the IETF as a solution to support mobility of IP terminal. In addition, SIP has been extended by research community to support terminal mobility. Both Mobile IP and extended SIP are compared for their ability to support mobility. It is found that Mobile IP is more suitable for supporting terminal mobility. However, Mobile IP is not sufficient for supporting intra-domain and high-speed mobility. An optimised solution therefore consists of Mobile IP and a specialised intra-domain mobility scheme. Having identified the characteristics and design principles of various existing inti-a-domain mobility protocols, a novel protocol known as Mobility-Aware Routing Protocol, MARP, is proposed. It is compared to some existing protocols and is proven to be superior in term of better handoff support, flexibility in packet routing and scalability. Mobile ad hoc networks are expected to be integral part of all-IP mobile network. Interworking of ad hoc and infrastructure-backed mobile networks enables extension of wireless networks coverage and provision of Internet access to mobile node with relaying. Key issues related to such a hybrid networking topology, in the context of ad hoc IP routing protocol and mobility management protocol, are identified. Mechanisms that allow interworking of mobility protocols and ad hoc routing protocols are designed. Interworking of both protocols significantly increases capability of ad hoc mobile nodes to interact with fixed network elements, but is at the cost of high control overhead. Effective techniques for reducing control overhead in the hybrid network is subsequently proposed.

Modern scientific research problems are getting more and more complicated. Addressing these problems require knowledge and expertise from a wide range of scientific disciplines. The instruments required for modern scientific research problems are also complex and expensive. In addition, the amount of research data generated by experiments on these problems is getter bigger to an extent that might not be manageable by any individual organisations. All of these factors have made global distributed collaborations become increasingly important in modern scientific research. Dealing with distributed collaborations at such a large scale has given rise to a new subject called e-Science. Grids have been widely accepted as promising infrastructures for e-Science. Grids enable the sharing of large-scale computational resources and experimental datasets in distributed virtual organisations. Web-based collaborative portals are commonly used as environments for interactions amongst distributed collaborators. Collaborators in a Web-based environment are subject to certain level of centralised administration and control. Their interactions have to be routed through a central server. This has been seen as inflexible and does not scale well with respect to the heterogeneity of distributed user communities. This thesis reports an investigation on a Collaborative e-Science Architecture (CeSA), which is an integration of Grid and Peer-to-Peer computing infrastructures using service oriented architecture, for supporting distributed scientific collaborations. CeSA leverages the advantages of Peer-to-Peer computing in supporting direct collaborations amongst end users and the capability of providing large-scale computational resources and experimental datasets. The investigation addressed two important issues with regard to the CeSA: (i) usability of the CeSA from users' point of view and (ii) an efficient resource discovery mechanism for the Peer-to-Peer environment. The usability was evaluated using the reaction kinetic research group in Leeds as a case study. An instance of the CeSA was prototyped for the evaluation. Feedback collected from the users was positive. An adaptive resource discovery approach has been introduced for the P2P collaborative environment of the CeSA. This adaptive approach takes into account the resource distribution and characteristics of scientific research communities. A learning mechanism, based on a classification of user interests using ontology, is used to adaptively route search queries to peers which are most likely to have the answers. Simulation results showed that this approach can efficiently improve query hit rates and also scale well with the increasing of network populations.

Ad hoc networks may be characterised as volatile systems of agents which collaborate to achieve a common goal, be that the fair use of a common pool resource, the minimisation of packet collisions or efficient data routing. Early approaches to these problems take inspiration from emergent systems, characterised in terms of a population governed by a fixed specification, and out of which arises a relatively sophisticated set of global properties. However, the features of open ad hoc networks necessitate an approach in which participating agents agree to conform to a mutable set of rules (permissions, powers and obligations), which are conventionally rather than physically binding. We are concerned with the development of open agent societies, in which a heterogeneous set of norm-aware agents with conflicting goals can produce similar emergent global properties by introspectively modelling system change. In this thesis we present norm-governed systems spaces, in which the `organised adaptation' of a specification is defined in terms of a finite number of adaptations and environmental states. These two elements are unified in a statistical framework which models an agent's motivation for system change as a set of Markov chains, mapped to a representation of utility. This arrangement allows participants to calculate the total expected utility of any deterministic adaptation strategy, wherein policy changes are chosen for the complete set of environmental fluctuations. We demonstrate this architecture in Matlab with a resource allocation example, in the context of which a norm-governed systems space is trained. Analytic models of agent societies are often idealised, and must be deployed in an experimental test bed to ensure the assumptions remain valid. We present the novel platform PreSage-NGS which constitutes a subsumption of the norm-governed systems space framework into a multi agent system simulation environment. We illustrate an application of this software in the same resource allocation scenario, identifying agent goal conflicts by inferring the alignment of private utility models through the evaluation of publicly expressed strategy proposals. This thesis explores how norm-governed systems spaces may be rapidly trained using lightweight approximations of agent populations in Matlab, which may be verified experimentally to ensure predictions translate to the reality of an open agent society. PreSage-NGS represents a means of exploring the strength of social models for system change in agent societies, in which bounded rationality may create unforeseen interaction dynamics.

Resource Management is a crucial aspect in the next generation cellular networks since the use of W-CDMA technology gives an inherent flexibility in managing the system capacity. The concept of a “Service Level Agreement” (SLA) also plays a very important role as it is the means to guarantee the quality of service provided to the customers in response to the level of service to which they have subscribed. Hence there is a need to introduce effective SLA-based policies as part of the radio resource management. This work proposes the application of intelligent agents in SLA-based control in resource management, especially when congestion occurs. The work demonstrates the ability of intelligent agents in improving and maintaining the quality of service to meet the required SLA as the congestion occurs. A particularly novel aspect of this work is the use of learning (here Case Based Reasoning) to predict the control strategies to be imposed. As the system environment changes, the most suitable policy will be implemented. When congestion occurs, the system either proposes the solution by recalling from experience (if the event is similar to what has been previously solved) or recalculates the solution from its knowledge (if the event is new). With this approach, the system performance will be monitored at all times and a suitable policy can be immediately applied as the system environment changes, resulting in maintaining the system quality of service.

Les progrès dans le domaine de la robotique mobile nous permettent aujourd'hui d'ajouter la notion de mobilité dans plusieurs classes de réseaux de capteurs sans fil. Le déploiement de capteurs mobiles est possible et utile dans de nombreuses applications (la sécurité, la surveillance de l'environnement, l’industrie et dans la santé). Deux thèmes sont abordées dans cette thèse: la conception d'un intergiciel pour les réseaux de robots mobiles et un ensemble de techniques pour le déploiement du robot mobile dans le cadre de réseaux de capteurs sans fil. L'intergiciel proposé et décrit dans cette thèse permet à l'utilisateur de facilement mettre en œuvre les différents types d'algorithmes de déploiement pour les robots mobiles. Dans la deuxième partie de la thèse, trois problèmes principaux sont présentés et analysés. Le premier est le problème de l'amélioration de la qualité de service avec l'utilisation des réseaux robotiques. Ce problème est résolu par l'utilisation de l'algorithme de déploiement qui améliore les performances de la transmission multimédia. Notre solution utilise une méthode intrusive pour recueillir la métrique de la qualité de transmission. Le deuxième problème abordé est la couverture de points d'intérêt avec des robots mobiles. Une station de base fixe est placé à l'intérieur du domaine d'intérêt, tandis que les robots mobiles disponibles couvrent le point d'intérêt et relayent l'information vers la station de base de proche en proche. Le troisième problème est la découverte des points d'intérêt et la couverture par l'utilisation de robots mobiles qui suivent des trajectoires circulaires concentriques afin de couvrir le domaine d'intérêt. En ajustant la vitesse de déplacement, ils répondent aux contraintes sur la couverture des points d'intérêt et la connectivité avec le station de base. / The advances in mobile robotics allow us today to add the mobility concept into many different classes of Wireless Sensor Networks. The deployment of mobile sensors is possible and useful in many application scenarios, ranging from the environmental monitoring and public safety applications, to the industry, healthcare and military applications. Two topics are elaborated in this thesis: networked robot middleware design and a set of approaches for mobile robot deployment in the context of wireless sensor networks. The middleware proposed and described in this thesis allows the user to easily implement different types of deployment algorithms for mobile robots. In the second part of the thesis, three main problems are presented and analyzed. The first is the problem of improving the quality of service with the use of mobile robotic networks, which is solved with the use of deployment algorithm that improves the performance of a multimedia communication by using an intrusive methods to gather the necessary transmission quality evaluation metrics. Second problem is the coverage of the point of interest with mobile robots, where the fixed base station is placed inside the field of interest, while the available mobile robots cover the point of interest and relay the information about it towards the base station in a multi-hop manner. The third problem is the point of interest discovery and coverage with the use of mobile robots which follow concentric circular paths to explore and cover the field of interest, and, by adjusting the movement velocity, they satisfy the constraints on PoI coverage and connectivity with the base station.

Avec l'essor de l'internet des objets, le nombre d'objets communicants dans nos villes est en pleine croissance, et notamment avec l'émergence du concept des ``Villes Intelligentes''. A l'instar des points d'accès Wifi, beaucoup de ces objets sont placés à des positions initialement connues, et pouvant être partagées dans le "Cloud" dans une démarche "BigData"; on parle alors de géo-référencement. Dans ce contexte, nous posons le problème de la radiolocalisation véhiculaire opportuniste basée sur la découverte de voisinage et l'exploitation des signaux reçus. Nous proposons une approche novatrice pour l'auto-localisation véhiculaire dans le domaine des systèmes de transport intelligents. La méthode proposée est opportuniste, passive et non-intrusive vis-a-vis des réseaux et des applications. La thèse défendue propose un système de localisation qui s'intègre aisément dans une pile de communication standard et qui s'étend sur deux niveaux protocolaires : couche physique et couche de liaison. Au niveau physique, nous utilisons un réseau d'antennes pour collecter les signaux à des fins d'identification et de localisation. Dès qu'un paquet est détecté au niveau MAC, les signaux physiques correspondants sont exploités pour estimer la position de l'émetteur dans le référentiel du véhicule. En utilisant l'identifiant de l’émetteur, on peut retrouver sa position dans un référentiel externe (ou global) depuis une base de données construite hors ligne et accessible localement ou à distance. En connaissant l'orientation du véhicule, la position du véhicule dans le référentiel externe peut être ensuite estimée par changement de base. / With the development of the internet of things, the number of communicating objects is rapidly increasing especially with the emergence of the "Smart City'' concept. As in the case of WiFi access points, many of these objects would operate at fixed and known positions. Such information is made available in the Cloud under the Big Data paradigm, thus leading to the so called geo-referencing of the nodes. In such a context, we pose the problem of opportunistic vehicular radio-localization by means of neighborhood discovery and signal processing of received neighboring signals. We propose a novel approach for vehicle self-localization as a contribution to the field of intelligent transportation systems. The proposed solution is opportunistic, passive and non-intrusive regarding the network operations and deployed applications. The thesis defends a localization system that seamlessly integrates with a standard communication stack and covers two layers: physical and link. At physical level, we use an antenna array for collecting and processing the received signals for both identification and localization. As soon as a packet is detected at MAC layer, the corresponding signals are exploited to localize the source with respect to the vehicle body frame. By using the source identifier, its position in the external (or global) frame is looked up in a database that was previously built off-line and made accessible locally or remotely. By assuming that the vehicle orientation is known, the position of the vehicle is then estimated using basis change equations.

Le Smart Grid (SG) permet une gestion intelligente du réseau électrique en mettant en place une infrastructure de communication pour l'échange des données. A travers cette thèse, nous avons proposé un protocole de routage des données pour les réseaux de capteurs sans fil (WSN) appliqués au niveau distribution des SGs. Nous avons proposé GRACO, un nouveau protocole de routage géographique basé sur l'intelligence d'essaim pour les WSNs. GRACO utilise le routage géographique comme mécanisme principal de routage de données, et un processus de recouvrement basé sur l'optimisation par colonies de fourmis pour contourner les vides de communication. Les performances du protocole ont été validées à travers des simulations. En comparant les résultats à l'état de l'art, le protocole proposé permet d'améliorer les performances du réseau en termes de taux de livraison, temps de réception et coût de livraison. Ensuite, nous avons proposé GRACO pour les réseaux de voisinages(NAN) sans fil du SG. GRACO permet d'assurer tous les types de paradigme de communication, surtout la communication Point-à-Point tout en offrant une évolutivité et des capacités d'auto-guérison. La faisabilité de GRACO dans les NANs a été confirmée à travers des simulations en utilisant des paramètres définis par le NIST.Enfin, nous nous sommes concentrés sur les qualités de service (QoS) requises par les applications du NAN. Nous avons ajouté un modèle de QoS à GRACO pour tenir compte de ces exigences. Les performances du protocole ont été validées à travers des simulations. Les résultats ont montré que le protocole permet de satisfaire les exigences en termes de fiabilité et de délai de réception des données. / The Smart Grid (SG) enables an intelligent management of the electrical grid. The implementation of SGs is conditional to the implementation of a communication infrastructure to exchange data between the entities connected to the grid. This thesis is positioned in the context of wireless sensor networks (WSN) in SGs. Through this work, we have proposed a data routing protocol for the communication network at the distribution level. First, we proposed GRACO, a new geographical routing protocol based on swarm intelligence for WSNs. GRACO uses the geographic routing as a main data routing mechanism, and a recovery process based on ant colony optimization to bypass communication voids. The protocol performances were validated through simulations. By comparing the results to the state of the art, the proposed protocol improves the network performances in terms of data delivery rate, end-to-end delay and delivery cost. Second, we proposed GRACO as the routing protocol for wireless neighborhood area networks (NANs) in SG. GRACO ensures all communication schemes, especially Point-to-Point communication while providing scalability and self-healing capabilities. The feasibility of the protocol in NANs was confirmed through simulations using parameters defined by NIST.Finally, we focused on the qualities of service (QoS) required by NAN's applications. We have added a QoS model to the proposed routing protocol to take account of NAN's communication requirements. The performances of the new protocol were validated through simulations. The results showed that the protocol can satisfy the most severe requirements in terms of reliability and end-to-end delay.

Réseaux de capteurs sont des réseaux composés de petits objets répartis dans l'espace, appelés nœuds ou capteurs, qui travaillent en collaboration - échange de messages sans fil - sur la même application. Aujourd'hui, ces types des réseaux sont largement utilisés dans le suivi environnemental, industriel et les applications grand public et à des fins militaires. Dans ces travaux, nous nous attaquons à différents domaines de recherche dans les réseaux de capteurs: contrôle de topologie, la mobilité, la découverte de voisinage et d'expérimentation à grande échelle. Nous utilisons une réduction de graphe des plus proches voisins avec les données obtenues d'alimentation du nœud pour développer l'algorithme de contrôle de topologie. Cet algorithme conserve une connectivité du réseau dans les situations critiques où certains des capteurs épuisent de leurs batteries. Les paramètres de découverte de voisinage sont utilisés pour en déduire la mobilité relative des capteurs. Ensuite, ces paramètres sont adaptés avec la puissance d’ émission pour obtenir un algorithme efficace de découverte de voisinage. Les sites d'expérimentation à grands échelle sont un outil précieux pour développer et tester des algorithmes pour les réseaux de capteurs sans fil, mais ils ont aussi des défauts divers, le plus grand d'entre eux est le coût. Nous présentons une émulation de réseaux à grande échelle comme une solution. On utilise de petits réseaux avec un placement précis des capteurs qui permet la réplication de comportement ainsi émuler des réseaux à grande échelle. Les algorithmes sont testés et évalués sur le simulateur WSNet et pratiquement en utilisant la plate-forme SensLab et nœuds de capteurs WSN430. / Wireless sensor networks can be defined as networks of small spatially distributed devices, called sensor nodes, which are working cooperatively - exchanging messages wirelessly - on the same application. Today these kinds of networks are widely used in environmental monitoring, industrial and consumer applications and for military purposes. In this thesis we are tackling different areas of research in wireless sensor networks: topology control, mobility, neighborhood discovery and large scale experimentation. We are using relative neighborhood graph reduction along with power supply data obtained from the sensor node to develop topology control algorithm. This algorithm maintains connectivity of the network in critical situations when some of the sensors drain their batteries. Neighborhood discovery parameters are used to deduce relative mobility of the sensor nodes. Then these parameters are adapted with transmission range to obtain energy efficient neighborhood discovery algorithm. Large scale experimentation sites are valuable tool for developing and testing of algorithms for wireless sensor networks but they also have various deficiencies, the biggest of them is cost. We present emulation of large scale networks as a solution. It uses small networks with the specific placement of the sensor nodes which allows replicating thus emulating behavior of the large scale networks. Algorithms are tested and evaluated on the WSNet simulator and practically using the SensLab platform and WSN430 sensor nodes.

En cas de sinistre, les infrastructures de communication peuvent être partiellement ou totalement détruites, ou inefficaces en raison du trafic élevé. Néanmoins, il est nécessaire d'assurer la connectivité entre les équipes de secours et le centre de commandement. Par conséquent, les solutions de communication temporaires sont essentielles jusqu'à ce que l'infrastructure soit rétablie. Dans cette thèse, nous nous concentrons sur le déploiement d’une solution de communication appelée réseaux de substitution. Ainsi, nous proposons un algorithme d'auto - déploiement pour permettre aux routeurs mobiles et répartis composant un réseau de substitution de couvrir la zone cible. Notre algorithme surveille les conditions du réseau pour décider si le routeur doit ou non se déplacer, ajustant la position de ce dernier en fonction des informations à un saut au moyen de la mesure active, c'est à dire, les paquets sondes. Ces paquets sondes permettent à l'algorithme de surveiller le canal et ses éventuels changements au fil du temps. Si le taux de transmission des paquets est suffisamment élevé, les connaissances obtenues seront exactes, cependant, le coût augmentera proportionnellement en consommant plus de ressources réseau. Par conséquent, nous proposons d'utiliser des données de substitution obtenus au moyen d'un estimateur autorégressif pour réduire la surcharge sans impacter notre algorithme de déploiement. Nous montrons par simulation l'efficacité des deux algorithmes et leurs performances en termes de temps de déploiement, de délai, de gigue et de débit. / In case of a disaster, the communication infrastructure may be partially or totally destroyed, or insufficient due to the high data traffic. Despite this, it is necessary to provide connectivity between the rescue teams and the command center. Therefore, temporary communication solutions are crucial until the infrastructure is restored. In this thesis, we focus on the deployment of communication solution called substitution networks. Thus, we propose a self-deployment algorithm to allow mobile routers that compose a substitution network spread out to cover the target area. Our algorithm monitors the network conditions to decide whether the router should move or not, adjusting its position based on one-hop information by means of active measurement, i.e., probe packets. Such probe packets allows the algorithm to monitor the channel and its eventual changes over time. If the probe transmission rate is enough high, the insights obtained will be accurate, however, the overhead will increase proportionally consuming network resources. Hence, we propose to use surrogate data obtained by means of an autoregressive estimator to reduce the overhead without impacting our deployment algorithm. We show by simulation the efficiency of both algorithms and their performance in terms of deployment time, delay, jitter, and throughput.