A study on estimation method of pedestrian's walking status for GPS positioning correction

Abstract

It is important to develop pedestrians, convenience and safety for realizing ITS world. In recent years, navigation for pedestrians has been increased by spreading of smart phones. The GPS (Global Positioning System) is main stream to obtain user's position information outdoors. However, positioning quality is poor in urban area. Then, the positioning reliability becomes worse. We can point out the following two problems. Firstly, the position information is shown on the sidewalk on the opposite side where the user is present although he does not cross the roadway. Secondly, despite the fact that he is not moving, the position information is shown as if the user is moving. Our purpose is to improve the above two problems. In this paper, we will propose two methods to estimate pedestrian's behaviors on roads. One is to estimate the time of starting road crossing. Another is to estimate pedestrian's state, that is moving or stopping. In this paper, we assume that the pedestrians is on the sidewalk of roads. Then, we will use a magnetic field sensor on the smartphone in order to estimate the above two pedestrian's behaviors. And we will show the ability of the both methods.