Indoor Wi-Fi tracking system using fingerprinting and Kalman filter

Abstract

Indoor tracking system is an important application to locate and track individuals and objects inside structures. Such systems are very helpful for visitors in campuses and large buildings, e.g. universities, hospitals and shopping malls. In this paper, we present and discuss the design of an Indoor Wi-Fi tracking system. One key advantage of the system is that it employs existing Wireless Local Area Network (WLAN) infrastructure. Using WLAN is attractive because it reduce the total cost of the overall system. The system consists of a mobile device running Android operating system and Wi-Fi Access Points (APs). For tracking and locating, the system applies K-Nearest Neighbor (KNN) based Wi-Fi fingerprinting method. To mitigate computation overheard, the system is partitioned to perform light weight computations at mobile device and execute heavy computations at server side. Several challenges addressed including AP accuracy handled by using stable APs and applying Kalman filter to reduce the signal fluctuation. Using Kalman filter improved the accuracy to 86% within 2m range of error.