Les Réseaux de Capteurs Sans Fils (RCSFs) ont pris beaucoup d'importance dans plusieurs domaines tels que l'industrie, l'armée, la pollution atmosphérique etc. Les capteurs sont alimentés par des batteries qui ne sont pas faciles à remplacer surtout dans les environnements peu accessibles. L'énergie de chaque capteur est considérée comme la source première d'augmentation de la durée de vie des RCSFs. Puisque la transmission de données est plus coûteuse en consommation d'énergie, notre préoccupation première est de proposer une technique efficace de transmission des données de tous les capteurs vers le sink tout en réduisant la consommation en énergie. Nous suggérons trois trois algorithmes d'agrégation de données basé sur la construction d'arbres : Depth-First Search Aggregation (DFSA), Flooding Aggregation (FA) et Well-Connected Dominating Set Aggregation (WCDSA) qui permettront de réduire le nombre de transmissions de chaque capteur vers le sink. L'agrégation des données basée sur la construction d'arbres souffre du délai de délivrance de données parce que les parents doivent attendre de recevoir les données de leurs feuilles. Certains parents pourraient avoir beaucoup de feuilles, et il serait alors assez coûteux pour un parent de stocker toutes les données entrantes dans sa mémoire. Ainsi, nous devons déterminer le temps que chaque parent doit mettre pour agréger et traiter les données de ses feuilles. Nous proposons un algorithme, Efficient Tree-based Aggregation and Processing Time (ETAPT) qui utilise la métrique Appropriate Data Aggregation and Processing Time (ADAPT). Etant donné la durée maximale acceptable, l'algorithme ETAPT prend en compte la position des parents, le nombre de feuilles et la profondeur de l'arbre pour calculer l'ADAPT optimal. A n'importe quel moment pendant l'agrégation des données par les parents, il peut arriver que la quantité de données collectées soit très grande et dépasse la quantité de stockage maximale de données que peut contenir leurs mémoires. Nous proposons l'introduction dans le réseau de plusieurs collecteurs de données appelés Mini-Sinks (MSs). Ces MSs sont mobiles et se déplacent selon un modèle de mobilité aléatoire dans le réseau pour maintenir la connexité afin d'assurer la collecte contrôlée des données basée sur le protocole de routage Mulipath Energy Conserving Routing Protocol (MECRP). Les capteurs peuvent être équipés de plusieurs interfaces radios partageant un seul canal sans fil avec lequel ils peuvent communiquer avec plusieurs voisins. La transmission des données à travers une liaison de communication entre deux parents peut interférer avec les transmissions d'autres liaisons si elles transmettent à travers le même canal. Nous avons besoin de savoir quel canal utiliser en présence de plusieurs canaux pour une transmission donnée. Nous proposons une méthode distribuée appelée: Well Connected Dominating Set Channel Assignement (WCDS-CA), pour calculer le nombre de canaux qui seront alloués à tous les capteurs de telle sorte que les capteurs adjacents se voient attribués des canaux différents / Wireless Sensor Networks (WSNs) have gained much attention in a large range of technical fields such as industrial, military, environmental monitoring etc. Sensors are powered by batteries, which are not easy to replace in harsh environments. The energy stored by each sensor is the greatest impediment for increasing WSN lifetime. Since data transmission consumes more energy, our major concern is how to efficiently transmit the data from all sensors towards a sink. We suggest three tree-based data aggregation algorithms: Depth-First Search Aggregation (DFSA), Flooding Aggregation (FA) and Well-Connected Dominating Set Aggregation (WCDSA) to reduce the number of transmissions from each sensor towards the sink. Tree-based data aggregation suffers from increased data delivery time because the parents must wait for the data from their leaves. Some parents might have many leaves, making it very expensive for a parent to store all incoming data in its buffer. We need to determine the aggregation time each parent in the tree has to spend in aggregating and processing the data from its leaves. We propose an Efficient Tree-based Aggregation and Processing Time (ETAPT) algorithm using Appropriate Data Aggregation and Processing Time (ADAPT) metric. Given the maximum acceptable latency, ETAPT's algorithm takes into account the position of parents, their number of leaves and the depth of the tree, in order to compute an optimal ADAPT time. At any time, the amount of data aggregated by parents may become greater than the amount of data that can be forwarded. We propose the introduction into the network of many data aggregators called Mini-Sinks (MSs). MSs are mobile and move according to a random mobility model inside the sensor field to maintain the fully-connected network in order to aggregate the data based on the controlled Multipath Energy Conserving Routing Protocol (MECRP). Sensors may use many radio interfaces sharing a single wireless channel, which they may use to communicate with several neighbours. Two sensors operating on the same wireless channel may interfere with each other during the transmission of data. We need to know which channel to use in the presence of multiple channels for a given transmission. We propose a distributed Well-Connected Dominating Set Channel Assignment (WCDS-CA) approach, in which the number of channels that are needed over all sensor nodes in the network in such a way that adjacent sensor nodes are assigned to distinct channels.

[pt] Os esquemas de alocação fixa de canais - FCA - conferem aos Sistemas de Comunicação Móveis Celulares estabilidade com o custo da necessidade de um pré-planejamento de freqüências, trazendo como conseqüência a baixa tolerância à variabilidade de tráfego. Algoritmos de alocação dinâmica de canais - DCA - tem sido propostos por diversos autores com o intuito de minimizar estes problemas, permitindo ao sistema flexibilidade no momento da escolha do canal candidato para servir a uma ligação. Esquemas de alocação híbrida de canais - HCA, combinam as técnicas de alocação fixa com alocação dinâmica de canais, conferindo ao mesmo tempo estabilidade e flexibilidade ao sistema.Este trabalho compara os algoritmos FCA, DCA e HCA sobre um mesmo cenário de simulação, permitindo analisar o desempenho dos mesmos. O esquema FCA utilizado é o mais simples, onde o primeiro canal com o nível aceitável de relação sinal interferência é escolhido para ser alocado. O esquema DCA utiliza a técnica de Segregação de Canais - CS, permitindo ao sistema maior flexibilidade no momento da escolha do canal candidato, devido a não existência do pré-planejamento de freqüências. O terceiro algoritmo, HCA, combina os dois esquemas anteriores. Ao final, será analisado o impacto de se priorizar o procedimento de handoff utilizando-se a técnica conhecida como Canais de Guarda. / [en] Fixed channel assignment -FCA- brings to the Cellular Communication Systems stability at the cost of the use of frequency planning, leading to low tolerance to traffic variability. Dynamic channel assignment -DCA- algorithms have been proposed by several authors in order to minimize these problems,incorporating flexibility to the system with respect to channel selection. Hybrid channel assignment - HCA- combines the techniques of fixed and dynamic channel assignment, giving to the system stability and flexibility at the same time. This work compares the FCA, DCA and HCA algorithms on the same simulation scenario, allowing a complete analysis of these approaches. The FCA used is the simplest. In this algorithm, the first channel with an acceptable level of signal to interference ratio is chosen to be allocated. The DCA uses the technique of Channel Segregation -CS-, a distributed self-learning algorithm that is shown to yield very good performance. The third algorithm,HCA, combine the two previous techniques. At the end, the impact of prioritizing the procedure of handoff will be analyzed, using the technique known as Guard Channel.