Urban Pedestrian Navigation Using Smartphone-Based Dead Reckoning and 3-D Map-Aided GNSS

This paper focuses on the pedestrian navigation in highly urbanized area, where a current smartphone and a commercial global navigation satellite system (GNSS) receiver perform poorly because of the reflection and blockage of GNSS signal by buildings and foliage. A 3-D map-aided pedestrian positioning method is previously developed to mitigate and correct the multipath GNSS signal. However, it still suffers from the low availability due to the insufficient number of satellites. We develop a smartphone-based pedestrian dead reckoning (PDR) algorithm, which is carried in the pedestrian's trousers. This PDR is capable of not only providing continues solutions but also indicating the pedestrian motions. A closed-loop Kalman filter with adaptive tuning is proposed to integrate the 3-D map-aided GNSS method with the smartphone-based PDR system. According to the experiment results, the proposed integration system can achieve ∼ 1.5-and 5.5-m of positioning errors in a middle-class and deep urban canyon, respectively.