|Abstract:||Distinct from wireless ad hoc networks, wireless sensor networks are data-centric, application-oriented, collaborative, and resource-constrained in nature. Especially, the limited resources on energy, bandwidth and computation capability radically change the considerations of system design.
In this thesis we propose a comprehensive framework of data-centric information processing and dissemination to realize the general mission for sensor networks, to extract useful information from the environment to users. Two main research issues are investigated in this thesis: (T1) coordination of sensors to generate valuable information from the sensed data and (T2) efficient dissemination methods to deliver the information of the best quality from sensors to subscribers. As to the first issue, we propose a self-organized, dynamic clustering approach for the target tracking system. Coordination between sensors is triggered by the events of interests and a cluster consisting of a leader and several sensors is formed dynamically. Through the probabilistic leader volunteering procedure and sensors replying method based on the quality of data, the proposed dynamic clustering algorithm effectively eliminates contention among sensors and renders more accurate estimates of target locations. Regarding the second issue, data dissemination consists of the problems of routing and transport layers. The problem of data transport in this thesis is formulated as a utility-based optimization problem, with the objective of maximizing the amount of information (utility) collected at sinks (subscribers), subject to both the channel bandwidth and energy constraints. Both the centralized and distributed approaches are devised to solve the optimization problem. To validate the design and to empirically study the performance of the proposed works, we implement a subset of the proposed works on the Motes testbed, including both the acoustic tracking system and the utility-based data transport components.