Abstract
The new generation of Android smartphones is equipped with GNSS chips capable of tracking multi-frequency and multi-constellation data. In this work, we evaluate the positioning performance and analyze the quality of observations collected by three recent smartphones, namely Xiaomi Mi 8, Xiaomi Mi 9, and Huawei P30 pro that take advantage of such chips. The analysis of the GNSS observation quality implies that the commonly employed elevation-dependent function is not optimal for smartphone GNSS observation weighting and suggests an application of the -dependent one. Regarding smartphone code signals on L5 and E5a frequency bands, we found that they are characterized with noticeably lower noise as compared to E1 and L1 ones. The single point positioning results confirm an improvement in the performance when the weights are a function of the -rather than those dependent on the satellite elevation and that a smartphone positioning with E5a code observations significantly outperforms that with E1 signals. The latter is expressed by a drop of the horizontal RMS from 8.44 m to 3.17 m for Galileo E1 and E5a solutions of Xiaomi Mi 9 P30, respectively. The best positioning accuracy of multi-GNSS single-frequency (L1/E1/B1/G1) solution was obtained by Huawei P30 with a horizontal RMS of 3.24 m. Xiaomi Mi 8 and Xiaomi Mi 9 show a horizontal RMS error of 4.14 m and 4.90 m, respectively.
Highlights
The ubiquity and high positioning performance of recent smartphones have widely expanded their primary application that was personal navigation [1]
These algorithms address the specific limitations of smartphone GNSS observations such as the low suppression to multipath and high observational noise highlighted by Riley et al [8], the carrier phase discontinuity that is driven by duty-cycle analyzed by Paziewski et al [9]
We analyze an accuracy level that may be reached with employed smartphones; we evaluate a potential benefit from the C/N0 -dependent weighting scheme and from the application of the signals that are transmitted on the second frequency band (L5/E5)
Summary
The ubiquity and high positioning performance of recent smartphones have widely expanded their primary application that was personal navigation [1]. The door to novel areas of market, industry, and science has been opened for smart, handheld, and low-cost GNSS devices [2,3,4]. A great deal of effort has been put into the development of observational and stochastic models that are suited to process smartphone GNSS observations [6,7] These algorithms address the specific limitations of smartphone GNSS observations such as the low suppression to multipath and high observational noise highlighted by Riley et al [8], the carrier phase discontinuity that is driven by duty-cycle analyzed by Paziewski et al [9]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
