A Smartphone-Based Online Pedestrian Positioning Approach for Both Structured and Open Indoor Spaces

Abstract

Pedestrian Dead Reckoning (PDR) technique has been extensively studied. However, without site survey, there are still some challenges in providing continuous and accurate online positioning services for indoor environments including both structured and open spaces, due to the lack of reference points. To cope with this problem, in this paper, we propose an online indoor pedestrian positioning approach based on integration of sensors in smartphone, indoor map and wireless access points (AP). In the structured spaces, we realize pedestrian positioning by matching activity sequence to indoor road map of structured spaces using a hidden Markov model. Meanwhile, we use Wi-Fi received signal strength (RSS) values to estimate AP locations. In the open spaces, with these APs’ location information, our approach applies an extended Kalman filter (EKF) model to integrate PDR and Wi-Fi positioning. We evaluate our approach using smartphones in the first floor of a teaching hall on campus. The results show that positioning errors of the proposed approach are decreased by $81.9$ percent in the structured space and $35.9$ percent in the open space when compared with PDR.