Sidewalk matching: a smartphone-based GNSS positioning technique for pedestrians in urban canyons

Abstract In urban canyons, most of Global Navigation Satellite System (GNSS) signals are blocked or reflected by high-rising buildings. The lack of redundant fault-free measurements poses a great challenge for detecting and isolating faulty measurements, degrading the GNSS positioning accuracy, particularly in the cross-street direction. The current GNSS positioning accuracy in urban canyons is not sufficient for identifying the sides of a street, which is highly demanded by numerous pedestrian applications. To address this issue, this study proposes a sidewalk matching technique to improve the redundancy of fault-free measurements and achieve positioning accuracy at sidewalk level in urban canyons. The technique relies on a simple pedestrian map and smartphone sensors, without Three-Dimensional (3D) building models. A new algorithm is proposed to identify the sides of the street, and a searching algorithm is developed to find GNSS position under constraints of the sidewalk. Extensive evaluation of the proposed technique based on a real-time platform demonstrated that the accuracy of GNSS positioning was less than 5 m, enabling seamless determination of the street sides in urban canyons. The new sidewalk matching technique can be used in various real-life applications, including pedestrian collision avoidance systems, online ride-hailing service, and jaywalking monitoring.