Design of a Practical WSN Based Fingerprint Localization System

摘要

Fingerprint positioning technology is among the most promising choices for seamless localization and is anticipated to be the future of seamless-locating services. The convenience of deployment and the high density signal source of wireless sensor networks (WSN) make them an ideal infrastructure for fingerprint positioning. In related researches, most WSN based fingerprint positioning systems are experimental demos that focus on the algorithm effectiveness and ignore the system reliability. This work proposes a practical WSN based fingerprint localization system. The system covers both indoor and outdoor scenarios and fulfills the demand for seamless localization. This paper work presents four measures that improve fault tolerance and system efficiency: a traffic regulation based radiomap (TRRM) establishing method, a full-overlapping clustering strategy, an adaptive feature space (AFS) algorithm, and a praxeological tracking algorithm. The proposed system is verified by hardware experiments on smart phones. Positioning accuracy is within 5 m in pedestrian tests and 10 m in driving tests.