| Advances in the design and development of low-cost, low-power-consumption sensors have led to active research in wireless ad-hoc sensor networks and their applications. Unlike other networks, sensor networks typically have restricted resources and limited power supply, which play a critical role in the design of sensor network architectures. Cross layer design approach is adopted to improve the network efficiency. This makes a generic information aggregation and dissemination architecture energy inefficient and even impractical considering sensor networks' wide range of envisioned applications. We propose and investigate all information aggregation and dissemination architecture to address three major problems in sensor networks: information dissemination, information aggregation and data fusion.;The problem of information dissemination concerns delivering information from the source node to all other nodes in the network. In this thesis, it is addressed in both static and mobile ad-hoc networks, specifically focusing on improving the broadcast efficiency and effectiveness. In static network scenarios, a protocol is proposed to dynamically adjust and improve the tree structure to decrease the overall broadcast cost. The proposed algorithm scales well and balances the broadcast load and power distribution across the sensor network. In mobile ad-hoc network scenarios, a constellation based scheme is proposed to provide flexible broadcast when relatively high node mobility is present. The scheme does not rely on any topology or neighborhood information. Its performance is within a constant of the global optimum. In order to retrieve information from the sensor network using mobile sinks, we design an information aggregation protocol to efficiently detect the shortest path between an arbitrary node to the sink as the sink is moving. Sensor nodes' low computational capabilities and limited energy motivate our design of a swarm intelligence based, energy aware protocol. SIMPLE, to route data to mobile destinations. SIMPLE prolongs the network's lifetime by evenly balancing residual energy across the network. Finally we present a data fusion scheme for sensor networks with a case study of target tracking to further improve the architecture's efficiency. |
