A Pedestrian Indoor Positioning System Based on the WI-FI and Walk Patterning Algorithm Using Mobile Device

Abstract

As the number of smart phone user is increasing, the demand for location-based services inside buildings is increased. Furthermore, inertial sensors are equipped in smart phone in these days. According to this tendency, many techniques for Location Based Service have been researched. This paper presents the navigation system consists of Pedestrian Walking Patterning (PWP) based on Pedestrian Dead Reckoning and Wi-Fi based fingerprinting method called Wi-Fi based Indoor Positioning System. PWP estimates step length and also includes the step detection with supplementation about error. In PWP, step length of user is estimated using the linear combination of acceleration, walking frequency and maximum gyroscope value. These parameters have linear pattern with step length. In addition to this, to reflect the walking characteristic of user, map data is used at step length estimation. WIPS estimates the position of user using Enhanced-kWNN. E-kWNN is the method to estimate the position of user using fingerprinting method. Existing fingerprinting methods use the fixed number of reference points to estimate the position. However, EkWNN varies the number of reference points to estimate the position according to a given environment. By varying the number of reference points, we could enhance the accuracy of positioning of user. Finally, we integrated PWP and WIPS as a navigation system and reduced the error caused by that algorithm using map matching algorithm. To verify the proposed navigation system, we did experiment in the hospital with lobby and hall and showed the result of positioning of user.