Edge-Centric Indoor-Outdoor Positioning Using LoRa Communication For Emergency Response
Abstract
Accurate localization is critical for emergency response in environments where communication infrastructure is limited or unavailable. Infrastructure-free methods such as Pedestrian Dead Reckoning (PDR) enable continuous positioning but suffer from drift, requiring multi-sensor fusion to maintain long-term accuracy. Existing systems typically offload this fusion to remote servers, which assumes reliable high-bandwidth connectivity that is rarely available in real emergencies. This paper proposes an indoor-outdoor localization system built on the combination of edge-centric fusion and LoRa communication to decouple positioning from network availability. A chest-mounted tag performs PDR estimation and streams PDR output, barometric, and GNSS data via BLE to a wearable edge device, which fuses these measurements with indoor map constraints in real time using a particle filter. Only compact 16-byte localization results are then transmitted via LoRaWAN to a monitoring system. Experiments in an indoor-outdoor environment demonstrate meter-level accuracy of ∼1.4m mean (1.78m RMSE) with 3.75s to 5.56s end-to-end latency.