Localization calibration using illuminance sensor for pedestrian dead reckoning with smartphones

Pedestrian dead reckoning is a method of localization that uses inertial sensors to estimate the travelled distance from the previous position. This method can estimate the position when GPS or WiFi is not available such as indoor. However, pedestrian dead reckoning is subject to cumulative errors. In this paper, we exploit the fact that most of lights are located regularly in indoor environments. We calibrate the estimated position by pedestrian dead reckoning by the estimated distance between lights when we judge that lights are located regularly from the information obtained from illuminance sensor. In our proposed method, we do not need to know the position of lights. We judge whether lights are regularly located from the change of illuminance obtained by illuminance sensor in smartphone. When we judge that lights are regularly located, we calibrate the estimated position by the average estimated distance between lights. Our experimental results show that the accuracy of localization with our proposed method is higher than that of the conventional one that uses only inertial sensors.