Optimal Adaptive Robust Kalman Filter for Smartphone Positioning

Anurag Raghuvanshi, Sunil Bisnath, Sudha Vana

Peer Reviewed

Abstract:	In the realm of Global Navigation Satellite System (GNSS)-based applications, smartphones equipped with multi-constellation and dual-frequency GNSS chipsets have demonstrated substantial potential. However, the inherent challenges associated with low-cost GNSS antennas and receivers necessitate innovative solutions to achieve higher accuracy. This paper explores an approach to improve the accuracy of smartphone positioning by employing measurement adaptiveness based on standardized residual of measurement residuals. A robust adaptive Kalman filter (RAKF) tightly-coupled (TC) GNSS Precise Point Positioning (PPP) and Inertial Measurement Unit (IMU) solution for smartphone is presented. In previous studies, the standardized residual of measurement innovations has been used as an effective statistic to form adaptive measurement noise covariances. The standardized residual of measurement innovations is effective in removing outliers and unmodelled errors. The standardized residual of measurement residuals is a good learning statistic (Hide et al., 2003a; Mohamed & Schwarz, 1999; Q. Zhang et al., 2018) to remove, or de-weight measurements affected by multipath. A RAKF utilizing t-test statistics to estimate the threshold for measurement adaptiveness is compared to a hybrid RAKF, which leverages the standardized residuals of measurement innovation and measurement residuals to mitigate the impact of abnormal measurements and utilizes position discrepancy statistics to mitigate effects of abnormal estimates. The study identifies an RAKF based on automatic thresholding as the optimal (best among methods discussed) solution with a reduction of 33% in 50th percentile errors and 23-34% reductions in 95th percentile error and 28-34% reductions in overall rms. The results are comparable to the RAKF based on standardized residuals of innovations.
Published in:	Proceedings of the 2024 International Technical Meeting of The Institute of Navigation January 23 - 25, 2024 Hyatt Regency Long Beach Long Beach, California
Pages:	25 - 41
Cite this article:	Raghuvanshi, Anurag, Bisnath, Sunil, Vana, Sudha, "Optimal Adaptive Robust Kalman Filter for Smartphone Positioning," Proceedings of the 2024 International Technical Meeting of The Institute of Navigation, Long Beach, California, January 2024, pp. 25-41. https://doi.org/10.33012/2024.19561
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In