Application of Adaptive Kalman Filtering on Smartphone Positioning

An Adaptive Kalman filter (AKF) is proposed which is used to estimate smartphone Global Navigation Satellite System (GNSS) pseudorange measurement variance. The filter is applied to stationary, bicycle and vehicle-based smartphone datasets collected in urban environments. The adaptive filter is compared to three other processing strategies: (i) conventional weighted least-squares, (ii) a velocity as random-walk Kalman filter (KF) for kinematic data or position as random-walk KF for static data, and (iii) an alternative KF implementation that uses Doppler to adapt process noise, all using a standard elevation and carrier-to-noise density ratio (C/N_0) measurement variance model. The adaptively estimated measurement variance is compared to the true error variance computed using the provided ground truth files and all four methods are evaluated in the position domain. The proposed AKF showed a horizontal positional accuracy improvement of 35.4%, 10.5%, and 27.3%, and a vertical positional accuracy improvement of 13.2%, 50.5%, and 59.6% for stationary, bicycle, and vehicle-based smartphone GNSS, respectively, compared to the second-best performing filter.