SENSOR: Data-Driven Sensor Networks Final Report (September 2003 to August 2006): Activities and Findings

In this project, we adopted a database approach to unite the seemingly conflicting requirements of scalability and flexibility in monitoring the physical world. The objective of this research was to build a new distributed data management layer that scales with the growth of sensor interconnectivity and computational power on the sensors over the next decades. Our system is designed to create the abstraction of a single processing node without centralizing data or computation. Since sensor networks are heavily resource-constrained, our research focused on cross-layer optimizations that are needed to exploit opportunities for the preservation of resources. Due to the regularity of query processing patterns we designed query-layer specific routing algorithms that are optimized — not for general point-to-point communication — but for the more regular types of communication patterns that are generated by a query layer. The main results of this project include a new MAC layer protocol that explores energy-delay tradeoffs, optimization techniques for processing multiple queries, a new approach to data aggregation based on the paradigm of universal optimization, new algorithms for efficient event notification in sensor networks, and a new algorithm for efficient data archival in sensor networks. The project also had an REU component in which we implemented a method for locating cars using a senso