More than 1 million Wireless Sensor Network (WSN) nodes have been deployed in 2005 alone, with an expected increase in deployment approaching 300%. Over 100 OEMs and service providers worldwide currently offer or are developing WSN products. For monitoring and control applications using WSN, the key enabling technology is the automatic localization of every sensor in the network. Sensor data must be correlated with the sensor’s physical location to permit the deployment of energy efficient routing schemes, source localization algorithms, and distributed compression techniques. Additionally, in inventory management and manufacturing logistics, localization and tracking of sensors is the primary purpose of WSN. Use of GPS is not feasible in large-scale networks, especially in cases involving low bandwidth applications. For WSN containing thousands, or even millions of nodes, the scale precludes any centralized location estimation. Thus a need exists for an automated, scalable algorithm for energy efficient sensor localization.
Researchers at the University of Michigan have developed a novel method for distributed calculation of low dimensional data from high dimensional data inputs for the purpose of radio location and data visualization. The method employs a scalable, distributed, multidimensional scaling algorithm that adaptively emphasizes the most accurate range measurements and naturally accounts for communication constraints within the sensor network. Each node adaptively chooses a neighborhood of sensors, updates its position estimate by minimizing a local cost function and then passes this update to neighboring sensors. For received signal strength based measurements, location estimates are nearly unbiased with variance close to the Cramér-Rao lower bound. Both received signal strength and time-of-arrival channel measurements are used to demonstrate performance.
Applications and Advantages
- Sensor localization and tracking in WSN
- Low power due to ability to calculate locations -nl-locally thus limiting transmission power
- Lower band-width requirements for WSN
- Decentralized processing eliminating the need -nl-for central-processing unit
- Non-increasing global cost of WSN
- Limited/reduced communication between sensors