CAIRN: Creating Anchors for Localization in Realistic Networks

Localization is a fundamental problem in wireless sensor networks. In many applications, sensor location information is critical for data processing and understanding. While the global positioning system (GPS) can be used to determine mote locations, the high cost often prohibits the ubiquitous use of GPS for location estimates. The cost of GPS has motivated researchers to develop localization protocols that determine mote locations using existing hardware and wireless measurements. Many of these protocols, however, have trade-offs and only perform well on a subset of networks and network configurations. We have developed a WSN localization algorithm that can be configured to perform well in a wide array of WSN configurations and deployments. We compare our algorithm to two well known and accurate algorithms: MDS-MAP and TSL. Through an extensive and credible analysis, we show that our algorithm, CAIRN, produces low localization error in a variety of network configurations. Additionally, we show how to manipulate algorithm parameters to tailor our algorithm for specific network configurations. Our work contributes an in-network WSN localization algorithm that produces accurate mote location estimates in various network configurations and we show that our algorithm outperforms the best existing protocols available.