Factor Graph Framework for Smartphone Indoor Localization: Integrating Data-Driven PDR and Wi-Fi RTT/RSS Ranging

Abstract

The classic fusion localization techniques based on the Kalman filter (KF) framework have been a focus of research community in the past decades, due to the limited computing power of mobile devices. However, with computing-efficient and sensor-rich smartphones now being commonplace, it is convenient and meaningful to provide more accurate positioning services for smartphones in an indoor environment. In this article, we design and develop a tightly coupled (TC) fusion platform of Wi-Fi round-trip time (RTT), received signal strength (RSS), and data-driven pedestrian dead reckoning (DPDR) based on factor graph optimization (FGO) for locating the consumer-grade smartphones in the indoor environment. Compared to the existing PDR solutions, including step model-based approaches and data-driven approaches, the proposed PDR solution with magnetic information (MI) constraint can track the relative position change of pedestrians at 20 Hz while supporting multiple smartphone usage poses. A comprehensive comparison between FGO, KF frame, and its variants is also performed. The experimental results demonstrate that the proposed fusion platform achieves an average positioning accuracy of 0.39 m. In addition, it also improves the accuracy of EFK and adaptive robust KF (ARKF) by 45.83% and 27.78%, respectively. The analysis shows that as the smartphone computing performance continues to improve, the FGO-based sensor fusion gradually replaces the KF frame and its variants.