Self-Contained Pedestrian Navigation Fusing ML-Selected GNSS Carrier Phase and Inertial Signals in Challenging Environments

The performance of the global navigation satellite system (GNSS)-based navigation is usually degraded in challenging environments, such as deep urban and light indoors. In such environments, the satellite visibility is reduced, and the complex propagation conditions perturb the GNSS signals with attenuation, refraction, and frequent reflection. This article presents a novel artificial intelligence (AI)-based approach, to tackle the complex GNSS positioning problems in deep urban, even light indoors. The new approach, called LIGHT, i.e., Light Indoor GNSS macHine-learning-based Time difference carrier phase, can select healthy GNSS carrier phase data for positioning, thanks to machine learning (ML). The selected carrier phase data are fed into a time difference carrier phase (TDCP)-based extended Kalman filter to estimate the user's velocity. Four trajectories including shopping mall, railway station, shipyard, as well as urban canyon scenarios over a 3.2-km total walking distance with a handheld device are tested. It is shown that at least half of the epochs are selected as usable for light indoor GNSS TDCP standalone positioning, and the accuracy of the velocity estimates can improve up to 88% in terms of the 75${\text{th}}$ percentile of the absolute horizontal velocity error compared with the state-of-the-art non-ML approach. Furthermore, a newly designed hybridization filter LIGHT-PDR that fuses the LIGHT algorithm and pedestrian dead reckoning solution is applied to perform seamless indoor/outdoor positioning in a more robust manner.