An energy-efficient architecture for wireless sensor networks
Networked embedded systems provide a versatile computing platform for supporting applications, such as environmental monitoring or military field surveillance. These networks are typically deployed in harsh environments and left unattended, which makes it impossible to re-charge their node batteries. Motivated by the large scale of sensor deployment, the high cost of wireless communications, the limited capabilities of sensor nodes, and the hostility of the encompassing environment, we design distributed protocols for topology management and network synchronization. We investigate problems that are specific to heavily-loaded sensor networks serving applications that can exploit data aggregation. We address applications that are either source-driven (in which sensors periodically send their reports to an observer), or query-driven (in which a subset of sensors respond to an observer's query). ^ We construct a hierarchical (clustered) network in which a subset of nodes are elected as super-nodes (or cluster heads). These heads constitute the routing infrastructure and aggregate data from their neighboring non-head nodes. The super-node functionality is rotated among network nodes to achieve load balancing and prolong the lifetime of every individual sensor. For sensors deployed in harsh environments, we propose a methodology for improving the connectivity of the hierarchical network and providing opportunities for multi-path routing. In addition, we consider applications and protocols that require node synchronization, and design a framework for time synchronization in multi-hop hierarchical networks that is based on relative synchronization. ^
Major Professor: Sonia Fahmy, Purdue University.