Enhanced Urban Localization Techniques Using GraphSLAM: Precision Improvements for Pedestrian and Underground Scenarios

Enhancing user localization offers a multitude of benefits, including heightened safety, precise navigation, improved transportation efficiency, and an overall enhanced user experience. It also enables personalized services, real-time information updates, and rapid emergency response. Traditional localization methods often rely on multiple sensors, such as inertial measurement units, radio frequency identification, ultrasound, or infrared sensors. However, deploying these sensors can be costly and challenging in various scenarios. Our paper aims to enhance user localization in different transport modes, particularly underground, where satellite signals are inaccessible. We leverage readily available resources such as Global Navigation Satellite System (GNSS) measurements when available, or else network measurements, transport-based Points of Interest (POIs), building corners, and doors. Our research demonstrates that incorporating transport-based POIs significantly improves the effectiveness of our approach. To evaluate our method, we conducted measurements, yielding promising results. Our approach achieved an average signal accuracy improvement of 1.7 meters for walking, underscoring its potential for substantial enhancements in user localization, especially for underground transport means.