Practical passive fingerprinting method

Abstract

Passive fingerprinting is a technique that collects the signal from a smartphone emitted from one location with a signal collection device called a Wi-Fi monitor, constructs a fingerprint with the received signal strengths of the collected signal, and uses it as a feature representing the user's location. Since the smartphone does not collect signals actively, the smartphone's sensor data cannot be used, so only the received signal strength emitted from the smartphone can be used for indoor positioning. However, since the signal strength is susceptible to fluctuations, inferring the user's walking path with only the received signal strength makes it challenging to construct a radiomap, an essential element for positioning. Besides, since the radiomap constructed by the passive fingerprinting method may cause a missing value problem and a shadow area problem due to a signal collection device, the radiomap's quality can be poor. In this paper, we proposed a method of constructing a passive fingerprinting radiomap using an adjacency matrix representing the WiFi monitor's deployment and an active fingerprinting radiomap. The proposed method reduces radiomap construction cost by filtering outlier monitors in the WiFi monitor sequence, indicating the user's path. Moreover, it solves the missing value problem and shadow area problem by combining the active fingerprinting radiomap with the passive fingerprinting method's radiomap.