Enhanced Smartphone Positioning in Urban Environments: GNSS Fault Detection and Mitigation Through Integrated Navigation System

Recent advancements in precise positioning, such as real-time kinematic (RTK) and precise point positioning (PPP), have demonstrated the potential to enable individuals to calculate accurate locations within the palm of their hands. Unlike conventional applications that primarily utilize smartphone location information for individual device use or cloud service connections, emerging applications are incorporating cooperative location concepts like machine-to-machine (M2M) and vehicle-to-vehicle (V2V). These services mainly rely on the absolute position derived from global navigation satellite system (GNSS) sensors embedded in smartphones. The performance of smartphone GNSS sensors, mostly provided by Google, typically offers positioning accuracy of approximately 3 to 5 meters under favorable conditions with an open-sky and observation of at least four satellites. However, these ideal conditions are not always met as many people use location-based services (LBS) on smartphones in urban areas with high-rise buildings and heavy vehicular presence, where usage is more frequent. In urban environments, GNSS signals are subject to interference from surrounding structures, leading to challenges such as multi-path errors where direct and reflected signals are received simultaneously, or situations where non-line-of-sight (NLOS) signals are received due to the blocking of direct signals. Additionally, smartphones, constrained by weight and cost considerations, often incorporate low-cost GNSS receivers and antennas, exacerbating these issues. Consequently, abnormal GNSS measurements received by smartphones significantly degrade the accuracy of user positioning and pose significant obstacles to cooperative-based location services. In this paper, we propose a solution to address these challenges by leveraging additional inertial navigation system (INS) sensors embedded in smartphones. In contrast to GNSS, which calculates the current absolute position based on the currently observed measurements, INS can predict the current position through integration based on the previous position. By comparing the currently measured GNSS data with the position predicted by INS, abnormal measurements can be detected and filtered out.