GPS World, January 2018

MOBILE PRECISION FIGURE 17 RTX convergence performance for a 35 hour dataset sampled on Nov 22 Broadcom chip at left and precision chip at right precision antenna processed by the engine that used the Broadcom chipset data that does not support L2 ranged between 6 and 10 satellites FIGURE 14 Convergence times were measured with post processing tools by splitting the datasets into individual time spans FIGURE 15 shows that the consumer GNSS chipset is able to get fixed ambiguity solutions but it takes considerably more time 266 seconds versus 4 seconds for the 95 of initializations However the system is fixing ambiguities and provides centimeter level positioning The same datasets were also processed with RTX Fast in California Thus the base station data was replaced by a global regional correction stream received from an internet based data source FIGURE 16 Horizontal accuracy for Broadcom reach 10 cm while the precision receiver reaches better than 3 cm The degradation is in part due to the difference in quality of the carrier phase and the different number of dual frequency satellites processed Precision devices provide measurements on E1 L1 L2 and L5 E5 providing at least dual frequency data from GPS GLONASS Galileo BeiDou and QZSS The Broadcom chipset tested provided dual frequency GPS and Galileo along with single frequency GLONASS and BeiDou however due to limited BeiDou constellation visible in California data from this constellation was not used Convergence was also analyzed and is shown in FIGURE 17 From the data we generated 24 convergence runs by taking one hour progressively shifting the start time by 5 minutes and running the data with different start times through the PPE engine This produced 24 runs which were translated into 68 and 95 convergence statics shown The RTX Fast solution for Broadcom reaches 30 cm horizontal error in 68 of the cases in approximately 12 minutes The RTXFast convergence using precision GNSS data is near instantaneous as can be seen in the right of Figures 16 and 17 reaching centimeter accuracy The code position solution using the RTX correction stream provides sub meter positioning FIGURE 18 As a summary the cumulative distribution function plots FIGURE 19 show the performance differences for 32 GPS WORLD WWW GPSWORLD COM JANUARY 2018 FIGURE 18 Code RTX performance for 35 hour dataset sampled Nov 22 and corresponding RTK and RTX phase solutions precision antenna this static environment on Nov 22 Cell Phone GNSS Antenna Results Similar tests were performed using an external cell phone GNSS antenna which is close to the antenna used in a typical smartphone RTK performance shows centimeterlevel accuracies and reasonable convergence times which are slightly worse than the results with the professional antenna FIGURES 20 24 In general as expected we achieve worse performance when connected to the GNSS cell phone antenna for all the different positioning modes For the cell antenna we also generated single frequency RTK and single frequency RTX Fast position solutions and compare it with a code positioning solution Positioning Engine in Android The results presented in this article captured GNSS data using the Android API and then postprocessed the data using PC versions of the position engines A significant amount of data has been captured and analyzed using this method For the purpose of real world demonstration the PPE has been implemented in an Android app to be used in cell phone devices This PPE is able to provide RTK RTX and code based positioning technology in one single PPE library The app has been tested running on a Samsung S7 FIGURE 19 CDF plots for the different PPE position solutions precision antenna