GPS PERFORMANCE EVALUATION OF THE HUAWEI MATE 9 WITH DIFFERENT ANTENNA CONFIGURATIONS

Transcription

1 GPS PERFORMANCE EVALUATION OF THE HUAWEI MATE 9 WITH DIFFERENT ANTENNA CONFIGURATIONS AND P10 IN THE FIELD Gérard Lachapelle & Research Team PLAN Group, University of Calgary ( GNSS Raw Measurements Taskforce Workshop GNSS Raw Measurements: From research to commercial use GSA Headquarters, Prague, 30 May 2018

2 Huawei Mate 9 Objectives (1) Antenna configurations: internal antenna One external antenna Two external antennas (spatial diversity) (2) No smartphone hardware modifications Scenarios: (a) simulator, (b) open-sky and (c) urban canyons Dynamics: static and vehicular kinematics Slides from Siddakatte, R., A. Broumandan and G. Lachapelle (2017) Performance evaluation of smartphone GNSS measurements with different antenna configurations. Royal Institute of Navigation International Navigation Conference, Brighton, U.K., November. 2/25 Lachapelle et al, 30May 2018

3 Data 1: Simulator tests (1/2) Code noise function of C/N 0 Signals generated by simulator fed to phone placed in a metallic box [to avoid spoofing nearby receivers] Static, no multipath/atmospheric errors, 15 min, controlled attenuation down to 25 db-hz C/N 0 3/25 Lachapelle et al, 30May 2018

4 Data 1: Simulator tests (2/2) Exposed cable end inside box acts as monopole antenna Google GNSSLogger App used to log raw measurements Pseudorange noise computed for each PRN using a fixed known position least-squares solution in two stages to smooth phone oscillator noise Code noise vs. C/N 0 : db-hz 5 30 db-hz m 25 db-hz 4/25 Lachapelle et al, 30May 2018

5 Data 2: Open sky tests Position & velocity with epoch-by-epoch least-squares East, North and Up position RMSE values: 6, 14 and 33 m Effect of noise, multipath and SV geometry 5/25 Lachapelle et al, 30May 2018

6 Data 3: Antenna configurations (1/2) Three configurations: (A) internal, (B) External helical, (C) External pinwheel External antennas: smartphone in metallic box to force phone to use these [and avoid spoofing of nearby receivers] A B C 6/25 Lachapelle et al, 30May 2018

7 Data 3: Antenna configurations (2/2) Mean C/N 0 values (db-hz) Pinwheel: Helical: Internal: LS positions obtained with raw measurements: Pinwheel Error Metric E N U RMS (m) Mean (m) Helical RMS (m) Mean (m) Internal RMS (m) Mean (m) /25 Lachapelle et al, 30May 2018

8 Data 4: Canyon tests (1/6) Huawei Mate9 measurements collected using 5 antenna configurations and phone locations A: Internal antenna, dashboard B: Internal antenna, phone in box inside car C: Internal antenna, car roof top D: External single helical antenna, roof top E: External dual helical antenna (spatial diversity), roof iphone 5SE also used to test its internal location accuracy with above five configurations 8/25 Lachapelle et al, 30May 2018

9 Data 4: Canyon tests (2/6) True trajectory obtained with NovAtel-SPAN LCI TM system (1 m) Highrise buildings with heights of 50 to 250 m along trajectory 9/25 Lachapelle et al, 30May 2018

10 Data 4: Signal quality (3/6) Five 10-min loops, one for each configuration Red dots (mean) & blue lines (standard dev) Mean gain of 6dB achieved with external antennas Better availability for spatial diversity case 10/25 Lachapelle et al, 30May 2018

11 Data 4: Multipath (4/6) computed using accumulated delta ranges (PLL) & pseudoranges Equivalent to Codeminus-carrier phase Only valid PLL delta ranges used Continuous carrier phase not possible due to dutycycle issue 10 to 150 m multipath 11/25 Lachapelle et al, 30May 2018

12 Data 4: HDOP & SV (5/6) B (high attenuationinside box) has poorest DOP & availability as expected C (roof) has better availability than A (dashboard) E (diversity) has the best availability and DOP 12/25 Lachapelle et al, 30May 2018

13 Data 4: Position & Velocity (6/6) Config. B was not providing LS solution for most trajectory (as expected) Lowest errors with external antennas Trajectory was eastwest hence northing results poorer LS position RMSE Case East North Up (m) (m) (m) Internal, dashboard Internal, box NA NA NA Internal, roof Ext. single helical Ext. dual helical LS velocity RMSE Case East North Up (m/s) (m/s) (m/s) Internal, dashboard Internal, box NA NA NA Internal, roof Ext. single helical Ext. dual helical /25 Lachapelle et al, 30May 2018

14 Data 4: Phone internal solutions Smartphone internal navigation solutions Huawei Mate9 & iphone 5SE: lowest errors with ext antenna Position RMSE (m) Configuration iphone 5SE Huawei Mate9 East North Up East North Up A. Internal, dashboard B. Internal, box C. Internal, roof D. External single helical D. External dual helical /25 Lachapelle et al, 30May 2018 Forward velocity RMSE (m/s) Configuration iphone 5SE Huawei Mate9 A. Internal, dashboard B. Internal, box C. Internal, roof D. External single helical E. External dual helical

15 Data 4: iphone 5S internal solution iphone 5S internal navigation solution 3D RMSE (m) A: 26.0, B: 37.5, C: 23.5, D: 16.5, E: 17.0 Position plots for iphone Ref A B C D E 15/25 Lachapelle et al, 30May 2018

16 Conclusions (Huawei Mate 9) C/N 0, pseudorange and position data confirm advantages of external antennas Main advantage of approach: (1) No smartphone hardware modification with use of external antenna (2) Use of metal container avoided spoofing issues Some remaining questions: (a) Indoor performance? (b) Two-antenna phase array advantage for antijamming? 16/25 Lachapelle et al, 30May 2018

17 Huawei P10 - Initial Field Testing {a} Open sky [clear horizon to 0 ] testing, low multipath, on a mountain top No Galileo data available on this {North American} model Reference coords: ITRF (approx WGS84) coords obtained with Trimble R10 geodetic receiver, single point [SP] and differential with base 40km away, GPS L1 data Static test with Huawei P10: 40 minutes P10 data processed with RTKLib epoch-by-epoch, single point (SP) and differential (DGPS), the latter with Trimble above R10 base. Use of RINEX files All data processed with Trimble R10 broadcast ephemeris 17/25 Lachapelle et al, 30May 2018

18 Trimble R10 Solution {b} As expected for such a geodetic quality receiver operating in L1 GPS mode with broadcast ephemeris 18/25 Lachapelle et al, 30May 2018

19 Trimble R10 Statistics {c} Mean differential carrier phase solution used to assess Huawei P10 solutions in next slides 19/25 Lachapelle et al, 30May 2018

20 Huawei P10 vs R10 {d} P10 data saved every 7-10 min, hence interruptions Carrier measurements initially good, then gradually degrading 20/25 Lachapelle et al, 30May 2018

21 P10 SP Vs R10 {e} 21/25 Lachapelle et al, 30May 2018

22 P10 DGPS Code Vs R10 {f} Statistics shown on next slide. Insufficient P10 carrier data to meaningfully evaluate performance of carrier phase solutions 22/25 Lachapelle et al, 30May 2018

23 P10 DGPS Code Vs R10 {g} Statistics shown on next slide. Insufficient P10 carrier data to meaningfully evaluate performance of carrier phase solutions 23/25 Lachapelle et al, 30May 2018

24 P10 vs R10 Statistics {h} Statistics shown on next slide. Insufficient P10 carrier data to meaningfully evaluate performance of carrier phase solutions 24/25 Lachapelle et al, 30May 2018

25 Initial Comments on P10 Not enough testing yet to fully evaluate unit Carrier phase performance not consistent; seems better at beginning of measurement sequence More lab testing needed as in the case of Mate 9 Seems difficult at this time to find a North American phone that also outputs raw Galileo data [e.g. Samsung S8] 25/25 Lachapelle et al, 30May 2018