The goal of this project is to develop a theoretical foundation for second generation sensor networks. There are two distinguishing features that are central to sensor networks. One distinguishing feature is that sensor networks are driven by sensor data, data which are not independent from node to node as in a data network, but which are correlated spatially as well as temporally. Thus one of the goals of the project is to investigate how to operate such networks efficiently taking full advantage of such correlations. A second distinguishing feature of sensor networks is that one is not generally interested in transporting all the sensory data generated at a node to another node. Rather, the purpose of a sensor network is to perform inference from the data (e.g., alarm networks) or make estimates from the data (e.g., mean temperature). More generally, the purpose of a sensor network is to compute an appropriate function of the data, both temporal as well as spatial, and make decisions based on the data. Thus the another focus of the research is an architecture for sensor networks to organize the in-network computation and communication taking advantage of all the distributed computational resources as well as the wireless communication medium, while respecting their possible limitations on energy, bandwidth, memory, etc. The tools used for conducting the research include network coding for distributed function computation, routing and data storage, distributed compression and cryptography. The results of the project will be disseminated through publications in major conferences and journals, and through a demonstration of data streaming for target monitoring and detection towards the end of the project.
