This is not available 017657

Wireless sensor networks (WSN) consists of a large number of smart sensor nodes. Each node can sense the data field around it, process the sensed data, and communicate it with the other nodes via wireless channels to accomplish an assigned common task. In most of its applications, there is always a data collection phase that involves sensing of the data at each sensor node and its subsequent transmission to the sink. Due to the limited battery power at each node, power efficiency is a critical design issue. The first part of the work in this dissertation is aimed at designing power- and delay-efficient data collection algorithms for cluster-based sensor networks. A novel transmission scheme bias been developed to support data collection in event-driven applications that are characterized by bursty traffic. The effect of Rayleigh fading channel within each cluster has been studied and a new MAC layer data transmission scheme has been proposed. Finally new MAC layer intra-cluster data retransmission algorithm has been proposed when the number of nodes within each cluster is unknown. The second part of the work has been aimed at the problem of real-time reconstruction of the sensed data field at a sink node. Assuming that the underlying field is spatio-temporally correlated, a distortion analysis framework is developed to analyze spatial and temporal distortion introduced by the limited number of sensor nodes and by the network delay associated with routing algorithms, respectively. The analysis framework is developed for a variety of WSN topologies, including cluster-based networks, heterogeneous networks, fixed topology regular networks such as linear arrays and ring networks, and random one-dimensional and two-dimensional grids.