Kalman Filter Based Integer Ambiguity. Ionosphere and Troposphere Estimation

Transcription

1 ION GNSS 2010 Kalman Filter Based Integer Ambiguity Resolution Strategy t for Long Baseline RTK with Ionosphere and Troposphere Estimation Tokyo University of Marine Science and Technology Tomoji jitakasu and Akio YASUDA

2 Contents Background Strategy for Long Baseline RTK Implementation Performance Evaluation Conclusion and Future Work 2

3 Background 3

4 RTK GPS/GNSS Real Time Kinematic GPS/GNSS cm level accuracy in real time Kinematic positions of moving receiver (rover) Carrier Phase Based Relative Positioning Transmit reference data to rover via wireless link Must resolve integer ambiguityon on the fly (OTF) Performance Depends on Baseline Length Reference Station Baseline Rover Receiver 4

5 Effect of Baseline Length E W BL=0.3 km E W BL=13.3 km U D U D N N S S BL=32.2 km BL=60.9 km U D U D N S N S E W E W (24 hr Kinematic : Fixed Solution : Float Solution) 5

6 Baseline Length and RTK Strategy BL (km) Error Elimination Ephem Ionos Tropos Others Strategy S 0 10 Broadcast M 10 L Conventional RTK 10 Dual Freq Broadcast 100 Interpolation Network RTK 100 1,000 VL >1,000 Real time Precise Dual Freq Estimate Long Earth Baseline ZTD + MF Tides (IGU) RTK Non RT Earth Post Estimate Precise Dual Freq Tides, Processing ZTD + MF (IGR, IGS) Ph WU or PPP 6

7 Application of Long Baseline RTK 100 km 1,000 km GPS Tsunami Monitoring System (Currently ~15 km off shore) 7

8 Strategy for Long Baseline RTK 8

9 Conventional AR Strategies Short Baseline RTK Rapid initialization/re initialization (OTF AR) Efficient integer vector search algorithm Ionosphere negligible in DD equations Medium/Long Baseline Post Processing Estimate WL/NL amb. by forming iono free LC Sequential rounding of WL/NL ambiguities Slow convergence Larger noise due to LC in kinematic i mode 9

10 Linear Combinations Coefficients Terms in DD Equation DD LC Noise Φ 1 Φ 2 P 1 P 2 R+T I 1 N 1 N 2 (cm) L1 λ λ L2 λ 2 1 γ λ P P2 1 1 γ 60 Notes 1 2 L3 C 1 λ 1 C 2 λ C 1 λ 1 C 2 λ Iono Free λ MW λ WL λ NL λ NL WL 0 0 λ LC /λ 1 /λ WL λ WL 42 2 (L1+P1)/2 λ 1 /2 1/2 1 0 λ 1 /2 30 Alt. Iono (L2+P2)/2 λ 2 /2 1/2 1 0 λ 2 /2 30 Free λ 1 =19cm, λ 2 =24cm, λ WL =86cm, λ NL =11cm, γ=f 12 /f 22, C 1 = γ/(γ 1), C 2 = 1/(γ 1) 10

11 AR Strategy for Long Baseline RTK No Linear Combination (LC) Use all original phase and code observables Not generate WL or NL LC Estimate ionosphere terms explicitly Suppress carrier phase noises Directly Resolve L1 and L2 Ambiguities (N 1, N 2 ) Search integer vector under ILS condition Efficient process by LAMBDA/MLAMBDA with shrinking space by linear transformation 11

12 Other Strategies Extended Kalman Filter for Real Time Est. Ephemeris IGU Predicted Orbit (Accuracy ~5 cm) Troposphere Estimate ZWD and gradient at ref and rover sites with mapping function (NMF) Other Corrections Receiver/satellite antenna PCV: IGS05.ATX Earth tides: solid earth tide model by IERS 12

13 Results by Simple Implementation : Fixed Solution : Float Solution AR Ratio Factor U D N S E W BL=300.0 km STD=2.2,2.3,3.2cm, Fix Ratio=46.4% 13

14 Partial Fixing of Ambiguities Search Type AR Strategy under ILS Condition All of ambiguities should be fixed at the same time Rising satellites often disturb ambiguity fixing Not tall Ambiguities iti must be Fixed Trade off between accuracy vs. fixing ratio Some Criteria to Determine Fixed or Float Variance of estimated ambiguity Duration of continuous valid data Satellite elevation angle 14

15 Feedback Fixed Ambiguity to Filter Open Loop AR Rover RTK Filter Ref. Float Solution Ambiguity Resolution Fixed Solution Fix and Hold Mode Rover Ref. RTK Filter Float Solution Integer Constraint for Fixed Ambiguities Fixed Solution Ambiguity it Resolution 15

16 Performance Improvement : Fixed Solution : Float Solution AR Ratio Factor U D N S E W BL=300.0 km STD=1.1,1.9,3.5cm, Fix Ratio=99.3% 16

17 Implementation 17

18 RTKLIB v241b v.2.4.1b Open source program package for GNSS positioning Whole source codes are freely available License: GPLv downloads (2.3.0) Portable library + several APs ANSI C + socket/pthread Portable command line APs GUI APs for Windows 18

19 RTKLIB APs on Windows RTKNAVI: Real time AP RTKPOST: Post Processing RTKPLOT: Plotting solutions RTKCONV: RINEX converter 19

20 Tests for Performance Evaluation 20

21 Offline Test BL= 29.9 km 1,099.9 km 1,000 km 2110 Tsukuba1 January1 7 7, 2009 (Winter) July 1 7, 2009 (Summer) 30 s x 1 week RINEX Rover: 477GEONET Stations Reference: GEONET 2110Tsukuba1 Ephemeris: IGU (predicted) Analysis S/W: RTKPOST v2.4.1b 21

22 Offline Test Results E W BL=471.2 km January 1 7, 2009 July 1 7, 2009 U D N S E W STD=0.7,0.9,2.3 cm FIX=99.8% BL=961.3 km STD=1.1,1.3,3.8 cm FIX=99.0% S N U D STD=1.6,1.3,3.0 cm FIX=98.8% STD=1.1,1.5,3.6 cm FIX=96.2% 22

23 Summary of Offline Test Results January 1 7, 2009 July 1 7, 2009 Ave=0.7,0.9,2.6cm Ave=1.4,1.6,5.2cm 0 0 Ave=97.8% Ave=93.4% Baseline Length (km) Baseline Length (km) 23

24 Real Time Test September 17 20, Hz x 72 H SUWN MIZU Rover: NovAtel BL= OEMV 3 + GPS 702 GG BL= km Reference: 1,024.8 km Rover IGS MIZU and SUWN MIZU SUWN BKG IGS IP Server NTRIP+RTCM v.3 CDDIS Receiver FTP (every 6H) RTKNAVI v.2.4.1b 24

25 Real Time Test Results (1) E W BL=435.7 km, STD=3.0,2.7,7.4cm, FIX=93.5% D U S N E W BL= km, STD=3.3,2.9,7.8cm, FIX=56.9% U D S N 25

26 Real Time Test Results (2) BL=435.7 km Ratio Factor Latency # of Sat BL= km 26 Ratio Factor Latency # of Sat

27 Conclusion and Future Work 27

28 Conclusion and Future Work A strategy for long baseline RTK proposed Offline test in 30 1,100 km baselines Real time test in 436 and 1,025 km baselines Proposed strategy works well up to 1,000 km baseline Performance degraded in summer time Need integration of meteorological info to improve troposphere correction 28