Automatic vehicle navigation is commonly used in today’s world with the proliferation of low-cost GNSS receivers. However, positioning inaccuracies inherent to GNSS-based positioning can reduce the application coverage. Areas such as urban canyons with limited satellite view and poor PDOP (position dilution of precision) could cause significant positioning errors. To complement the inefficacy of positioning in such
urban environment, we conducted a study into the effectiveness of using WIFI fingerprinting technology to complement GNSS-based positioning for vehicle positioning in an urban environment. WIFI fingerprinting positioning complements well with GNSS-based positioning in that although there is limited satellite view in urban environment, there are an abundance of WIFI signals. In this paper, we present the results of our study
conducted in the Singapore Jurong Lake District. Data collection of WIFI signals were performed using an Android APP running on a mobile device by driving around the area. The WIFI signal data are then processed and stored in a WIFI fingerprint database. A navigation APP was developed to compute the vehicle position based on measured WIFI signals and WIFI fingerprint data from the database. The vehicle position can then
be displayed on a map live, just like a GNSS-based navigation device. We studied two methods of fingerprinting-based positioning. The first method uses the kNN method in which WIFI signal strengths information is taken into the positioning computation. The second method uses a simpler approach without using signal strengths. The positioning accuracies obtained using these two methods of WIFI fingerprinting are
discussed. The results showed that WIFI fingerprinting-based positioning may be a feasible complement to GNSS in providing positioning for vehicle navigation in an urban environment.