1 of 2404/21/23

GNSS-based Flight Inspection

Systems

International Flight Inspection International Flight Inspection SymposiumSymposium

Oklahoma City, OK USA June 2008Oklahoma City, OK USA June 2008

Euiho KimSelex-si-us

Todd Walter, David Powell Stanford University

2 of 2404/21/23

Introduction• Stanford Univ. has developed Global

Navigation Satellite System (GNSS)-based flight inspection systems:– WAAS-aided FIS: described at 14th IFIS at Toulouse – WAAS-based FIS– Standalone (unaugmented) GPS-based FIS

• The primary objective of this work was to develop a FI truth system with– High efficiency: onboard self-contained FIS– Low cost: no INS (WAAS-based FIS, standalone

GPS-based FIS)

3 of 2404/21/23

WAAS(SBAS)-based FIS• Sensors

– WAAS receiver (single frequency), Radar altimeter (RA), TVPS

• Positioning strategy− Reference position @ runway threshold from the RA & TVPS + a

relative positioning algorithm, called “T-D PRP”, with GPS/WAAS

• Characteristics – On-board self contained system– Does not use an INS– Does not need to fly over the entire runway– Near real-time positioning (WAAS optimized)– A receiver must be able to output raw GPS/WAAS measurements

• Reference – “WAAS-based Flight Inspection System”, Journal of Aircraft, , Vol.

45, No 2, AIAA

4 of 2404/21/23

EstimatedTrue

Trajectory

T-D PRP

RA,TVPS

Ref. Position at threshold

Raw GPS/WAAS• carrier phase meas.• pseudo-range• ephemeris• WAAS messages

WAASPosition

Fast-ClockCorrection

Filtering (MV)

Satellite Exclusion Tests

Y/N

Y/NSat 1

Y/NSat 2

…Ref. Position Validation

Y

Y

WAAS-based FIS Architecture

∑

Approach

RunwayThreshold

Approach finishes

Relativeposition

5 of 2404/21/23

Time-Difference Precise Relative Positioning (T-D PRP)

• Ranging source− Time-difference of carrier phase measurements− Fine resolution (~cm)− Immune to any bias-like GPS ranging error sources

• Utilizes WAAS correction messages if available − Does not use WAAS ionospheric delay corrections

• Reduce ionospheric effects− Ionospheric delay gradient is estimated and removed using the

combination of code and carrier (near real time)− The estimation process is further checked by exclusion tests

• Improvements− Fine resolution relative position (~cm)− Not vulnerable to abnormal ionospheric effects

6 of 2404/21/23

Satellite Exclusion Tests

GPS/WAASSatellite Health

Designation

Cycle-Slip

Residual Test on Estimated Ionospheric

Gradient

SatelliteOutage

• Satellites and carrier phase measurements must pass the following checks

7 of 2404/21/23

Reference Position Validation

T-D PRP

True

WAAS

Bias

Off-Set

Range wrt. Threshold

VerticalPosition

RadarAltimeter

• Test statistics for the reference position validation using WAAS position in vertical

• WAAS 95% accuracy (WAAS PAN Oct ~ Dec, 2006)

1.11 m in horizontal

1.34 m in vertical

• RA and TVPS 95% Accuracy

RA: 15cm

TVPS: 15cm

thr(WAAS T-D PRP) (RA T-D PRP )| |

| WAAS vertical bias Error (RA) |

VT AVG

<Vertical position characteristics for WAAS, RA, T-D PRP>

8 of 2404/21/23-10000

-5000

05000

-1

0

1

2

x 104

0

200

400

600

800

1000

EAST (m)NORTH (m)

UP

(m

)

Flight Test Data• Flight test during Oct 30~31, 2006 in collaboration with

FAA flight inspection crews• 23 approaches used

–No radar altimeter or TVPS data was available, unfortunately• Truth source is RTK DGPS

9 of 2404/21/23

WAAS-based FIS Test Results• Assumes a perfect reference position

– Radar altimeter and TVPS were not available– A reference position was obtained from RTK DGPS

• Graphs show only the T-D PRP algorithm errors– (Total error = ref. position bias (RA, TVPS) + T-D PRP error)

<Horizontal> <Vertical>

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)

Cro

ss-T

rack

Err

or(

m) CAT I

Error Limit

CAT II&III Error Limit

CAT II&III Error Limit

CAT I Error Limit

10 of 2404/21/23

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)

Cro

ss-T

rack

Err

or(

m)

WAAS-based FIS Accuracy (95%) at Critical Region

• Error analysis− T-D PRP error grows over time− Total errors at critical region

= Ref. position bias + T-D PRP error @ CR

• T-D PRP error statistics @ CR

• Therefore, with:– RA accuracy (95%) : 15 cm– TVPS accuracy (95%) : 15 cm– T-D PRP Accuracy @ CR

• Total Error @ CR is:– Up : 18 cm < 30 cm (req.)– Cross-track : 17 cm < 60 cm (req.)

Up (m) Cross-Track (m)

Mean -0.04 0.00

Std 0.05 0.04

RMS 0.06 0.04

CR

11 of 2404/21/23

Standalone GPS-based FIS• Sensors

– GPS receiver (single frequency), Radar altimeter, TVPS

• Positioning strategy− Reference position @ runway threshold (RA, TVPS) + relative

positioning using the T-D PRP algorithm with standalone GPS

• Characteristics– On-board self contained system– Does not use INS– Does not need to fly over the entire runway– Near real time positioning– Operational worldwide – A receiver must be able to output raw GPS measurements

• Reference– “Standalone GPS-based Flight Inspection System”, ENC-GNSS 2007,

Geneva, Switzerland

12 of 2404/21/23

Standalone GPS-based FIS Architecture

EstimatedTrue

Trajectory

T-D PRP

RA,TVPS

Ref. Position at threshold

GPS measurements• carrier• pseudo-range• ephemeris

Satellite Exclusion Tests

Y/N

Y/NSat 1

Y/NSat 2

…FIS-RAIM

Y

∑

Approach

RunwayThreshold

Approach finishes

13 of 2404/21/23

Integrity Issues• The WAAS-based FI system derived its

integrity from the underlying WAAS integrity messages.

• GPS-based FI system does not have that feature; therefore,– An autonomous procedure was developed to

meet that need– Uses redundant satellite measurements to

establish integrity of the FI procedure

14 of 2404/21/23

Receiver Autonomous Integrity Monitoring (RAIM)

• Basic RAIM methods to detect satellite failures− Range comparison method [Lee, 1986]− Least-squares-residual method [Parkinson and Axelrad, 1988]− Parity method [Sturza, 1989]− Maximum separation of solutions [Brown and McBurney, 1987]

Sat 1 Sat 5

Sat 2 Sat 4

Sat 3unhealthy

nominalfault

<Illustration of separated solutions under 1 satellite failure>

d

15 of 2404/21/23

Flight Inspection System RAIM (1/3) (FIS-RAIM)

• Current baseline RAIM scheme – Not suited for high-accuracy

application [Ober,06]– Detects large satellite failures

• Minor satellite failures – Slow ramp clock error (a few cm/s)– Low amplitude satellite clock

dithering (a few meters)

• FIS-RAIM– Designed to detect the violation of

CAT I error limit

0 2000 4000 6000 8000 10000 12000-6

-4

-2

0

2

4

6

Range wrt. Threshold (m)

UP

Err

or(

m)

ramp clock error (4cm/s)

CAT I Error Limit

CAT II & IIIError Limit

0 2000 4000 6000 8000 10000 12000-6

-4

-2

0

2

4

6

Range wrt. Threshold (m)

UP

Err

or (m

)

sinusoidal clock dithering (1m amp.)

CAT I Error Limit

CAT II & IIIError Limit

16 of 2404/21/23

0 2000 4000 6000 8000 10000 120000

1

2

3

4

Range wrt. Threshold (m)

Max

imum

Sep

arat

ion

of S

olut

ions

(m

)

0 2000 4000 6000 8000 10000 120000

2

4

6

8

10

Range wrt. Threshold (m)

Max

imum

Sep

arat

ion

of S

olut

ions

(m

)

• FIS-RAIM

– Uses maximum separation of solutions of T-D PRP in vertical

– Uses full measurements during approach

– Uses 2 test statistics • A slope of observed

maximum separated solutions using linear regression (ramp clock error)

• RSS for the residual test on the regression (clock dithering)

sat. clock ramp error(4cm/s)

sat. clock dithering (1m amp.)

Flight Inspection System RAIM (2/3)(FIS-RAIM)

17 of 2404/21/23

0 2000 4000 6000 8000 10000 120000

1

2

3

4

5

6

7

Range wrt. Threshold (m)

Max

imum

Sep

arat

ion

of S

olut

ions

(m

)

Threshold

• The thresholds are– Slope: 0.05/2 deg (determined for CAT I error limit)– RSS : (20cm)2*chi2inv(99.9%,n-2)

no satellite failures

Flight Inspection System RAIM (3/3)(FIS-RAIM)

0 2000 4000 6000 8000 10000 12000-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

Range wrt. Threshold (m)

Res

idua

ls (

m)

no satellite failures

18 of 2404/21/23

Standalone GPS-based FIS Test Results

• Assumes a perfect reference position– Radar altimeter and TVPS were not available – The reference position was determined from DGPS

• Graphs show only T-D PRP errors– (Total error = ref. position bias (RA, TVPS) + T-D PRP error)

<Vertical> <Horizontal>

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)

Cro

ss-T

rack

Err

or(m

)

CAT I Error Limit

CAT II & III Error LimitCAT I

Error Limit

CAT II & IIIError Limit

19 of 2404/21/23

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)C

ross

-Tra

ck E

rror

(m)

Standalone GPS-based FISAccuracy (95%) at Critical Region

(CR)• Error analysis

− T-D PRP error grows over time− Total errors at critical region

= Ref. position bias + T-D PRP error @ CR

• T-D PRP error statistics @ CR

• Therefore, with:• RA accuracy (95%) : 15 cm• TVPS accuracy (95%) : 15 cm• T-D PRP Accuracy @ CR

• Total Error @ CR is:– Up : 18 cm < 30 cm (req.)– Cross-track : 17 cm < 60 cm (req.)

Up (m) Cross-Track (m)

Mean -0.04 0.00

Std 0.05 0.04

RMS 0.06 0.04CR

20 of 2404/21/23

WAAS-based FIS vs. GPS-based FIS

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)

Cro

ss-T

rack

Err

or(

m)

0 2000 4000 6000 8000 10000 12000-3

-2

-1

0

1

2

3

Range wrt. Threshold (m)

Cro

ss-T

rack

Err

or(m

)

WAAS-based FIS

WAAS-based FIS

GPS-based FIS

GPS-based FIS

• Almost identical accuracy− Strong temporal correlation

in satellite clock-ephemeris residual correction error of WAAS and GPS (usual)

− A FI approach occurs over a very short time

21 of 2404/21/23

Summary of WAAS-based FIS

• WAAS-based FIS Accuracy (95%)– Meets flight inspection system accuracy requirement for CAT

III ILS calibration • Up : 18 cm < 30 cm (req.) @ CR• Cross-track : 17 cm < 60 cm (req.) @CR

• Benefits– Firm integrity

• Satellite health from WAAS integrity messages • Protect against sharp and nonlinear ionospheric delay gradient• Reference position validation

– Low cost and high efficiency

• Limitation– Operational where WAAS (or any SBAS) is available

22 of 2404/21/23

Summary of Standalone GPS-based FIS

• Standalone GPS-based FIS accuracy (95%)– Meets flight inspection system accuracy requirements for

CAT III ILS calibration • Vertical : 18 cm < 30cm (req.) @ CR• Cross-track : 17 cm < 60cm (req.) @CR

• Advantages– Firm integrity

• Satellite health from FIS-RAIM• Protect against sharp and nonlinear ionospheric delay gradient

– Low cost and high efficiency for worldwide use

• Limitation– Not able to check the integrity of a reference position

23 of 2404/21/23

Accuracy Efficiency Cost-Effect. Worldwide Usefulness

Inertial-based AFIS

DGPS-based AFIS

WAAS(SBAS)-based FIS

Standalone GPS-based

FIS

Comparison with Current FIS’s

24 of 2404/21/23

Thank you !Special thanks to FAA AVN for the generous

support in this research