Crowd-assisted radio map construction for wi-fi positioning systems

Abstract

Numerous attempts have been made to estimate position in an indoor space where a GPS signal is not available or contains errors. To date, the use of a radio map, which collects the received signal strength indicator (RSSI) of Wi-Fi in advance, is the least expensive and easiest method. However, to construct such a radio map, it is necessary to visit the site and collect signals. This work is not only labor intensive but also requires periodic updates due to changes in the Wi-Fi environment. Hence, the original purpose of using radio maps to reduce costs has not yet been fully satisfied. This paper proposes a practical method for constructing Wi-Fi radio maps in a large shopping mall environment. In the first step, the initial radio map is constructed based on a small number of reference data from the mobile payment system. In the second step, the path of the crowd-sourced fingerprint sequence is estimated and labeled to improve the radio map. The crowd-sourced fingerprints was collected from commercial location service that is actually being used.