Integration of GNSS and cellular infrastructure for large-scale indoor localization

Wireless indoor localization systems remain among the most promising technologies for enabling location-based services. Nevertheless, the inherent constraints of indoor environments often necessitate dense infrastructure deployment, presenting significant barriers for broader adoption due to the associated costs of installation and maintenance—particularly in vast areas such as underground parking facilities and metro stations. This paper introduces a novel localization framework that extends Global Navigation Satellite System (GNSS) signals indoors by repurposing existing passive Distributed Antenna Systems (DAS), originally designed for cellular communication. Within this framework, the distributed antennas acts as GNSS anchors by retransmitting satellite signals. Upon tracking these retransmitted signals and acquiring raw GNSS measurements, indoor pseudorange residuals and rates are extracted to infer both user location and velocity. A Particle Filter (PF)-based localization algorithm is then proposed for data fusion, the combination of anchor identification and ranging with regional constraints facilitates an efficient particle initialization, while inertial cues derived from GNSS Doppler further support consistent motion updates. Additionally, subregion information and anchor handover events offer essential spatial context for dynamically refining particle weights. Field experiments using a commercial smartphone in an actual underground parking environment demonstrate the system’s ability to deliver high localization accuracy and scalable coverage with minimal additional infrastructure.