Time-bounded essential localization for wireless sensor networks
IEEE/ACM Transactions on Networking (TON)
| Hi-index | 0.00 |
Sensor networks are seen as one of the most promising emerging technologies. In many cases, sensed information is useless without knowing the sensor location. In addition, many of the networking protocols currently being developed for sensor networks assume the availability of node positions. In the first part of this dissertation, we provide a theoretical foundation for the internode distance-based localization problem (fine-grained localization). We construct "grounded graphs" to model network localization and apply graph rigidity theory to develop conditions for the unique localizability of networks as well as to construct and characterize classes of uniquely localizable networks. In addition, we study the computational complexity of network localization and prove that the problem is NP-complete, even for uniquely localizable networks where a distance measurement exists between every two nodes within a certain distance r of one other. We then describe subclasses of uniquely localizable grounded graphs for which localization can be efficiently computed. In the remainder of our work, we study localizability and localization in sensor networks of randomly-placed nodes and present new algorithms for controlled mobility, deployment and localization. Our localization algorithm succeeds in unambiguously localizing a much wider class of networks than was previous possible. In random networks, almost all localizable nodes are localized using our approach, even at node densities close to the minimum required for localizability. We close with an example application of localization in a novel controlled mobility algorithm for the optimization of power usage in sensor networks.