L1 SFMC SBAS System to Improve the Position Accuracy of Android Device

Abstract

Since Google launched the Android N in 2016, more than 30 brand new Android devices are providing user-domain with GNSS raw measurements, which are pseudorange, carrier phase and doppler observables. Before the Android N there was no other way to improve the position accuracy of the smart phone or tablet than A-GPS or DGPS-CP, which were only available at server-side or under a constraint of satellite visibility. Even though there is not yet any GNSS augmentation functionality implemented to feed the correction to the chipset, the access to the raw measurements allow users to improve the position accuracy of the devices and to applies them to various areas as desired. There are typical ways to correct the device pseudorange-based position such as DGNSS (Differential GNSS), PPP (Precise Point Positioning), SBAS (Satellite Based Augmentation System) and so on. In the case of DGNSS, DGNSS reference stations broadcast the PRC (Pseudo Range Correction) whose service area is 150km to 200km from each station, thus the user should include a function of selecting the reference station according to the user’s location. PPP is another augmentation solution which provides a global precise positioning service that uses detailed physical models and corrections, and precise GNSS orbit and clock products. One of the many advantages of PPP is that the PPP solutions of GNSS orbit and clock products are global. Currently, the IGS (International GNSS Service) provides users with the RTS products which consist of GNSS satellite orbit and clock corrections to the broadcast ephemeris in real time, whose data rate in IGS RTS (Real Time Service) product stream is about 2,000bps. SBAS is also a powerful enhancement to augment GPS by broadcasting augmentation messages including differential corrections and integrity information with a very low data rate of 250bps from a GEO satellite. Unlike DGNSS, all the users in the SBAS service area receive the same correction and information, so the user app does not need any model for selecting the reference station. Moreover, the data bandwidth is only 250 bps, which is about 10% of the PPP service, thus the users does not need to feel burdened by the cost of mobile data communication even though current GNSS module in Android device should receive the SBAS message on the internet, not on the L1 signal. Despite of the simplicity at the user side and the low data rate for the correction message, current SBAS system cannot fully utilize the multi-constellation effect of the Android devices. All the current SBAS systems except of the SDCM of Russia provide GPS-only augmentation messages, thus when Android smart devices supporting multi-constellations(GPS, GLONASS, Galileo, etc.) cannot utilize the advantages(i.e. improving the accuracy) of multi-constellations when applying the current SBAS message. To complement this disadvantage, we propose the L1 SFMC (Single-Frequency, Multi-Constellation) SBAS in this paper. The L1 SFMC SBAS method and its message protocols were the method suggested at ION GNSS+ 2017 to improve the performance of the L1 single frequency GNSS by containing all the information for the multi-constellation GNSS in the modified SBAS message format. We propose to remove the fast corrections and to assign additional time slots for providing the long-term corrections of GLONASS and other constellations. To include our suggested L1 SFMC SBAS in the suggested message format, we generated the long-term corrections using the precise orbit and clock products from the IGS Data Center. Also, we generated the ionospheric corrections for all the constellation using IGP(Ionospheric Grid Point) vertical delay estimates which has already been included in SBAS Message Type 26 and IONEX data. We carried out some experiments to confirm the effect of our suggested L1 SFMC SBAS corrections on Android smart device at Seoul, Korea. As a result of applying the L1 SFMC SBAS corrections to Android GNSS raw measurements on Samsung Galaxy S8 with 250bps, the mean values have been reduced from 2.02m(horizontal) and 9.49m(vertical) to 0.36m and 0.49m, and the RMS errors have been reduced from 5.19m(horizontal) and 10.56m(vertical) to 1.30m and 2.06m.