Efficient data collection in sensor networks
Subhash Suri, Divyakant Agrawal, Chiranjeeb Buragohain · 2006
Wireless sensor nodes are tiny self-contained devices which combine sensing, computation and radio communication in a single package powered by batteries. An ad hoc network of such devices can allow efficient environmental monitoring over a large geographic area. The primary characteristic of these sensor networks is their size (many thousands of nodes) and limited resources of individual nodes, especially battery power (couple of AA batteries). Since radio communication is the largest consumer of battery power, the utility and lifetime of such a sensor network is limited by the communication load in the network. The constraints arising from scalability and limited power require that we have to fundamentally rethink the protocols and algorithms that are deployed over such a network. Since the sensors continuously monitor physical variables, they generate large quantities of data. Users of the sensor network make queries over this network with the data of interest spread over thousands of nodes over a large geographic area. Thus the key problems in sensor networks arise in gathering this data and processing it efficiently to answer queries. In this dissertation we focus on designing efficient data collection algorithms which ensure reliable delivery of data while minimizing resource consumption over the network. Building upon these foundations, we design algorithms which can carry out efficient data processing within a single node, as well as over the entire network.