Utilising Raw GNSS Data from Android Units to Improve Positioning

Abstract: In this study it was examined if positioning can be improved by utilising raw Global Navigation Satellite System (GNSS) data from Android devices instead of using Android's existing Location API. If this is the case, the coworkers at BM System may modify their apps, one is called Loco, to develop one that calculates positions directly by utilising raw GNSS data instead of using Android's Location API which is done presently. There is an app called GNSS Compare that does just this, estimates positions with raw GNSS data. In addition, if the phone using the app has dual frequency capabilities it uses this for better calculations by eliminating one of the errors, the ionosphere delay, that occur in positioning. To allow evaluation of the accuracy of the positions, the app had to be altered to write to .gpx files. In order to test the accuracy, field tests were performed. These were executed by going to certain points, geodetic control points, which have a precise position which is known prior and testing the accuracy. Two phones were included in this study, Xiaomi Mi 8 and Samsung Galaxy S9 with the former having dual frequency capabilities and two apps were similarly tested, GNSS Compare and BM's app Loco. At each control point, Xiaomi Mi 8 with GNSS Compare, Xiaomi Mi 8 with Loco, Samsung Galaxy S9 with GNSS Compare and Samsung Galaxy S9 with Loco were tested statically. These points were then compared by evaluating the median squared error from the control point as well as the minimum squared error. It was concluded that Xiaomi Mi 8 with GNSS Compare performed the best out of the four combinations of phones and apps with an average of its median values being 6.20 meters. It was also the second best when comparing the average minimum values, closely following Xiaomi Mi 8 with Loco with the former scoring at 1.27 meters and the Xiaomi with Loco 99 cm. The lowest best value for Xiaomi Mi 8 with GNSS Compare was 5 mm, which is a very small number, and as a result was the best combination of phone and app in that aspect. In conclusion, Xiaomi Mi 8 with GNSS Compare seems to produce the highest accuracy of positioning amongst the phones and apps tested. A reason behind this is that the app calculates the positions by utilising the raw data of the phone compared to Loco which uses the Android Location API. In addition, Xiaomi M 8 is a phone with dual frequency capabilities and since GNSS Compare uses this property when applying their ionosphere delay correction it may give this combination an advantage compared to the others. However, this needs to be further studied.