end-to-end performance analysis of the paris in-orbit demonstrator ocean altimeter
DESCRIPTION
END-TO-END PERFORMANCE ANALYSIS OF THE PARIS IN-ORBIT DEMONSTRATOR OCEAN ALTIMETER. Salvatore D’Addio, Manuel Martin-Neira , Chris Buck Acknowledgment: Nicolas Floury, Roberto Pietro Cerdeira. TEC-ETP, Electrical Engineering Department European Space Agency. The PARIS Concept. PARIS = - PowerPoint PPT PresentationTRANSCRIPT
IGARSS-2011, 25-29 July, Vancouver (Canada)
END-TO-END PERFORMANCE ANALYSISEND-TO-END PERFORMANCE ANALYSISOF THE OF THE
PARIS IN-ORBIT DEMONSTRATOR PARIS IN-ORBIT DEMONSTRATOR OCEAN ALTIMETEROCEAN ALTIMETER
Salvatore D’Addio,Salvatore D’Addio, Manuel Martin-NeiraManuel Martin-Neira, Chris Buck, Chris Buck
Acknowledgment: Nicolas Floury, Roberto Pietro Acknowledgment: Nicolas Floury, Roberto Pietro CerdeiraCerdeira
TEC-ETP, Electrical Engineering DepartmentTEC-ETP, Electrical Engineering DepartmentEuropean Space AgencyEuropean Space Agency
IGARSS-2011, 25-29 July, Vancouver (Canada)
The PARIS ConceptThe PARIS Concept
PARIS = PARIS =
Passive Reflectometry and Interferometry SystemPassive Reflectometry and Interferometry System
IGARSS-2011, 25-29 July, Vancouver (Canada)
PARIS IOD Key FeaturesPARIS IOD Key Features• Direct Cross-correlation (DxR)Direct Cross-correlation (DxR)
– High gain beams for direct signals (D)High gain beams for direct signals (D)– High gain beams for reflected signals (R)High gain beams for reflected signals (R)– Implicit use of full GNSS bandwidth (3x40 Implicit use of full GNSS bandwidth (3x40
MHz)MHz)• Dual or Trial Frequency for precise estimation of Dual or Trial Frequency for precise estimation of
ionospheric delay ionospheric delay • Precise on-board delay calibrationPrecise on-board delay calibration• Precise on-board amplitude calibrationPrecise on-board amplitude calibration
L1L5 L2
40 MHz 40 MHz 40 MHz
IGARSS-2011, 25-29 July, Vancouver (Canada)
PARIS IoD L1 Power WaveformPARIS IoD L1 Power Waveform• GPS L1 Composite, Orbit Height: 500Km, Down-Looking GPS L1 Composite, Orbit Height: 500Km, Down-Looking
Antenna Gain: Antenna Gain: 19dBi, 19dBi, Nadir Looking GeometryNadir Looking Geometry
IGARSS-2011, 25-29 July, Vancouver (Canada)
PARIS IoD Mission Summary TablePARIS IoD Mission Summary TableInstrument PARIS Altimeter
Main Scientific Product Mesoscale Ocean Altimetry
Secondary Scientific Product
Polar Ice Thickness
Scientific by-products
Ionospheric total electron contentWind speed and wind direction over oceanSignificant Wave HeightMean Square Slope of ocean surfaceSea Ice ExtentRiver and lake water level
Particular Application Tsunami detection
In-orbit Demonstrator
Operational Mission
Orbit Polar Sun
SynchronousPolar Sun
Synchronous
Orbital Height 800 km 1500 km
Swath 900 km 1500 km
Revisit Time 3 days 2 days
Spatial Resolution 100 km < 100 km
Antenna Diameter / Gain 0.9 m / 19 dB 2.4 m / 30 dB
Number of Beams 4 16
FrequenciesGPS L1+L5
Galileo E1+E5
All 3 bands ofGPS, GLONASS,
GALILEO, BEIDOU
Total Altimetry Accuracy (1)
< 17-20 cm < 5-7.5 cm
Platform PROBA-like PROTEUS-like
Launcher Configuration Piggy back Main passenger
IGARSS-2011, 25-29 July, Vancouver (Canada)
Factors affecting altimetry Factors affecting altimetry performanceperformance
Factors affecting Factors affecting altimetry performancealtimetry performance
CommentComment
PrecisionPrecision Allowed incidence angle of specular point
• Higher incidence angle reduces the height precision, but increases swath• Scanning reduces antenna gain
Instrument Characteristics:1. Antenna Gain, Receiver NF 2. XC Coherent integration
time
1. Improve SNR2. Optimized for maximum
precision
Waveform Retracking Performance
Optimum retracker to be defined
AccuracyAccuracy Uncertainties on Propagation:1. Ionosphere2. Troposphere3. EM Bias4. Skewness Bias
1. Use of 2 or 3 Frequencies2. Estimated or measured3. Frequency Dependent4. Fraction of EM Bias
Uncertainties on Geometry/Orbit
Residual instrument errors after calibration
On-board Delay/Amplitude Calibration needed
IGARSS-2011, 25-29 July, Vancouver (Canada)
Max Incidence Angle (1/2)Max Incidence Angle (1/2)
P
G
O
s
Ionosphere
i i
The incidence angle plays an important role in the definition of The incidence angle plays an important role in the definition of critical parameters of the PARIS mission, such as:critical parameters of the PARIS mission, such as:
(a) Precision(a) Precision
(b) Sampling and Coverage(b) Sampling and Coverage
IGARSS-2011, 25-29 July, Vancouver (Canada)
Max Incidence Angle (2/2)Max Incidence Angle (2/2)
a) Altimetric performance degrades with incidence anglea) Altimetric performance degrades with incidence angle
b) Ionospheric delay and antenna loss increase with incidence b) Ionospheric delay and antenna loss increase with incidence angleangle
The incidence angle should be restricted (<35The incidence angle should be restricted (<35).).Coverage and sampling obtained by raising orbital height. Coverage and sampling obtained by raising orbital height.
(a(a)) (b)(b)
dh
P
G
O
s
Ionosphere
i i
i
ddh
cos2
i
d
Antenna gain
IGARSS-2011, 25-29 July, Vancouver (Canada)
Factors affecting altimetry Factors affecting altimetry performanceperformance
Factors affecting Factors affecting altimetry performancealtimetry performance
CommentComment
PrecisionPrecision Allowed incidence angle of specular point
• Higher incidence angle reduces the height precision, but increases swath• Scanning reduces antenna gain
Instrument Characteristics:1. Antenna Gain, Receiver NF 2. XC Coherent integration
time
1. Improve SNR2. Optimized for maximum
precision
Waveform Retracking Performance
Optimum retracker to be defined
AccuracyAccuracy Uncertainties on Propagation:1. Ionosphere2. Troposphere3. EM Bias4. Skewness Bias
1. Use of 2 or 3 Frequencies2. Estimated or measured3. Frequency Dependent4. Fraction of EM Bias
Uncertainties on Geometry/Orbit
Residual instrument errors after calibration
On-board Delay/Amplitude Calibration needed
IGARSS-2011, 25-29 July, Vancouver (Canada)
Interferometric Processing SNRInterferometric Processing SNR
• Up/Down-looking Antenna Gain and Up/Down-looking Antenna Gain and Up/Down RX NF shall be such to ensure sufficient Up/Down RX NF shall be such to ensure sufficient SNRSNR
• In PARIS IoD with interferometric processing, the In PARIS IoD with interferometric processing, the final power waveform SNR is given as:final power waveform SNR is given as:
int
11
cr
R
D
SNRSNR
SNR
SNR
IGARSS-2011, 25-29 July, Vancouver (Canada)
Interferometric Processing SchematicInterferometric Processing Schematic
DELAYDOPP SHIFT
XC
DOCONBPF
fo
DOCONBPF
fo
cT
dt
UPCHAIN
DOWNCHAIN
|.|2W2(τ)
W1(τ)
W3(τ)
WN(τ)
iN
W fs
Ts
For a given target along track spatial resolution, e.g. For a given target along track spatial resolution, e.g. 100Km100Km, , the coherent integration time and the incoherent averaging the coherent integration time and the incoherent averaging are are inversely proportionalinversely proportional: :
Tc Ninc
100 Km c incT T N
Cross-Correlation Integration Time Cross-Correlation Integration Time (Tc)(Tc)
IGARSS-2011, 25-29 July, Vancouver (Canada)
t SP
P
t
Power Altimetry Waveform
1 chip delay
22
, ,
,
sin cS R
c
fTZ P ACF
fT
Pulse limited footprint
θs
φs
P=(θ,φ) SP
Iso-Doppler
RX PARISsatellite
TX GNSSsatellite
direct path “d”
Specularreflected path
Generic reflected path “r”
Iso-Delay
Cross-Correlation Integration Time Cross-Correlation Integration Time (Tc)(Tc)
IGARSS-2011, 25-29 July, Vancouver (Canada)
22
, ,
,
sin cS R
c
fTZ P ACF
fT
PA
RIS
Flig
ht D
ir
RX
Pulse Limited
FootprintDoppler FootrpintFor Low Tc
3dB Antenna Footprint
Low TcLow TcDoppler
Limited
Footprint
PA
RIS
Flig
ht D
ir
RX
Doppler FootrpintFor High Tc
3dB Antenna Footprint
High High TcTc
Cross-Correlation Integration Time Cross-Correlation Integration Time (Tc)(Tc)
IGARSS-2011, 25-29 July, Vancouver (Canada)
2
, ,
,
sin cSP R c
c
fTSNR P ACF T
fT
SNRSP
TC
PA
RIS
Flig
ht D
ir
RX
PA
RIS
Flig
ht D
ir
RX
PA
RIS
Flig
ht D
ir
RX
3dB Antenna Footprint
PA
RIS
Flig
ht D
ir
RX
Cross-Correlation Integration Time Cross-Correlation Integration Time (Tc)(Tc)
IGARSS-2011, 25-29 July, Vancouver (Canada)
Cross-Correlation Integration Time Cross-Correlation Integration Time (Tc)(Tc)
2 2,
, ,
1 1 11
2cos
Z S
hinc
inc SP Z S
c P
SNR SNRNP
cSNR T100 Kminc
c
TN
T
2 2
12 2
1 11
h c
c c
const Tconst T const T
• Analytical model for altimetry precision:Analytical model for altimetry precision:
• Tc shall be chosen such to have a sufficient SNR (e.g. 6dB) but cannot be increased further otherwise the altimetric precision is degraded because speckle is not averaged enoughTc shall be chosen such to have a sufficient SNR (e.g. 6dB) but cannot be increased further otherwise the altimetric precision is degraded because speckle is not averaged enough• However, Tc cannot be too low otherwise consecutive XC samples may be correlated and speckle would anyway not be averagedHowever, Tc cannot be too low otherwise consecutive XC samples may be correlated and speckle would anyway not be averaged
IGARSS-2011, 25-29 July, Vancouver (Canada)
Factors affecting altimetry Factors affecting altimetry performanceperformance
Factors affecting Factors affecting altimetry performancealtimetry performance
CommentComment
PrecisionPrecision Allowed incidence angle of specular point
• Higher incidence angle reduces the height precision, but increases swath• Scanning reduce antenna gain
Instrument Characteristics:1. Antenna Gain, Receiver NF 2. XC Coherent integration
time
1. Improve SNR2. Optimized for maximum
precision
Waveform Retracking Performance
Optimum retracker to be defined
AccuracyAccuracy Uncertainties on Propagation:1. Ionosphere2. Troposphere3. EM Bias4. Skewness Bias
1. Use of 2 or 3 Frequencies2. Estimated or measured3. Frequency Dependent4. Fraction of EM Bias
Uncertainties on Geometry/Orbit
Residual instrument errors after calibration
On-board Delay/Amplitude Calibration needed
IGARSS-2011, 25-29 July, Vancouver (Canada)
Waveform RetrackingWaveform Retracking
• The retracking has the objective of estimating the sea state The retracking has the objective of estimating the sea state parameters and the sea mean surface height by fitting a parameters and the sea mean surface height by fitting a theoretical model to measured waveformstheoretical model to measured waveforms
• Maximum Likelihood Estimation (MLE) or on weighted least Maximum Likelihood Estimation (MLE) or on weighted least squares estimation has been extensively used in Conventional squares estimation has been extensively used in Conventional Radar Altimetry.Radar Altimetry.
• The MLE method estimates the parameters by determining which The MLE method estimates the parameters by determining which values maximize the probability of obtaining the recorded values maximize the probability of obtaining the recorded waveform shape in the presence of noise of a given statistical waveform shape in the presence of noise of a given statistical distribution. distribution. 0( ) , , spW f SWH
SWH
delay
IGARSS-2011, 25-29 July, Vancouver (Canada)
Factors affecting altimetry Factors affecting altimetry performanceperformance
Factors affecting Factors affecting altimetry performancealtimetry performance
CommentComment
PrecisionPrecision Allowed incidence angle of specular point
• Higher incidence angle reduces the height precision, but increases swath• Scanning reduce antenna gain
Instrument Characteristics:1. Antenna Gain, Receiver NF 2. XC Coherent integration
time
1. Improve SNR2. Optimized for maximum
precision
Waveform Retracking Performance
Optimum retracker to be defined
AccuracyAccuracy Uncertainties on Propagation:1. Ionosphere2. Troposphere3. EM Bias4. Skewness Bias
1. Use of 2 or 3 Frequencies2. Estimated or measured3. Frequency Dependent4. Fraction of EM Bias
Uncertainties on Geometry/Orbit
Residual instrument errors after calibration
On-board Delay/Amplitude Calibration needed
IGARSS-2011, 25-29 July, Vancouver (Canada)
The impact of IonosphereThe impact of Ionosphere
• At L-band, the ionosphere is a major contributor to the At L-band, the ionosphere is a major contributor to the propagation propagation delay delay Iih cos2
TECf
I2
3.40
21
1 cos2f
Iih
25
5 cos2f
Iih
iih
cos226.2
cos226.1 15
15
59.215
h
withwith
• Multi-frequency observations allow to remove the Multi-frequency observations allow to remove the ionospheric effectionospheric effect
• However this is at the expense of a severe error amplification factorHowever this is at the expense of a severe error amplification factor
IGARSS-2011, 25-29 July, Vancouver (Canada)
Possible Ionospheric Correction Possible Ionospheric Correction Method Method
• Multi-frequency observations are taken:Multi-frequency observations are taken:
• A regression is performed over N samplesA regression is performed over N samples of ionospheric delay of ionospheric delay
•The smoothed ionospheric delay is used in the range equations, The smoothed ionospheric delay is used in the range equations, resulting in an improved error amplification factor resulting in an improved error amplification factor
x
h
f
f
y
y2
5
21
2
1
1
1
• The ionospheric delays are retrieved with the following The ionospheric delays are retrieved with the following uncertainties:uncertainties:
yxf 78.1
12
1
yxf 2.3
12
5
yxNf
78.112
1
yxNf
2.312
5
yh NN
4
2.3
4
78.1
2
1 22
2
xy h
f
22cos N
xy h
i f
512
1
1.26 2.262cos 2cos
x
f i i
2 21 515 1 5
1 1
2 2cos 2 2cosN
Ih f f
i i
151.27 for 3 (linear regression)h y N
1251.08 for 5 (4th order regression)h y N
• Starting from range equation:Starting from range equation:
IGARSS-2011, 25-29 July, Vancouver (Canada)
Ionospheric Correction: Example Ionospheric Correction: Example 1/21/2
• What matters in mesoscale ocean altimetry is the difference What matters in mesoscale ocean altimetry is the difference between between the ionospheric delay at two specular points the ionospheric delay at two specular points Worst case is between nadir (i=0 Worst case is between nadir (i=0) and edge of swath (i=30) and edge of swath (i=30))
P
G2
O
s2
Ionosphere
i i
G1
s1
IGARSS-2011, 25-29 July, Vancouver (Canada)
Ionospheric Correction: Example Ionospheric Correction: Example 2/22/2
0.000
2.000
4.000
6.000
8.000
10.000
12.000
-8000.0 -6000.0 -4000.0 -2000.0 0.0 2000.0 4000.0 6000.0
-10
-8
-6
-4
-2
0
2
4
6
-8000.0 -6000.0 -4000.0 -2000.0 0.0 2000.0 4000.0 6000.0
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
-8000.0 -6000.0 -4000.0 -2000.0 0.0 2000.0 4000.0 6000.0
Vertical Delay (m)Vertical Delay (m)
Mesoscale Delay (m)Mesoscale Delay (m)
Residual Delay (m)Residual Delay (m)
200 km averaging of mesoscale delay applied200 km averaging of mesoscale delay applied 0.15 m0.15 m
+0.15 m+0.15 m
+6 m+6 m
10 m10 m
+12 m+12 m
0 m0 m
Derived from RA-2 real dataDerived from RA-2 real data
IGARSS-2011, 25-29 July, Vancouver (Canada)
PARIS IoD Preliminary Error BudgetPARIS IoD Preliminary Error Budget
ParameterIOD Height Total Accuracy on 100Km, G=20dBi, H=800Km
Instrument Noise and Speckle 12.5 cm
Ionosphere Averaging Noise 9.5 cm (2 frequencies,
N=3)
Ionosphere Residual 5 cm
Troposphere (Wet and Dry) 5 cm
EM Bias 2 cm
Skewness Bias 1 cm
Orbit / Geometry 5 cm
Instrument error residuals 2 cm
Total RMS Height Accuracy 18 cm at Edge of Swath
IGARSS-2011, 25-29 July, Vancouver (Canada)
CONCLUSIONSCONCLUSIONS
• The PARIS altimeter can provide synoptic views of the ocean being thus The PARIS altimeter can provide synoptic views of the ocean being thus well suited for mesoscale ocean altimetrywell suited for mesoscale ocean altimetry
• The interferometric processing between direct and reflected signals The interferometric processing between direct and reflected signals maximises the altimetric performance of the systemmaximises the altimetric performance of the system
• A demonstration mission should be able to achieve 18 cm total accuracyA demonstration mission should be able to achieve 18 cm total accuracy
• An operational mission could be targeted for 5 cmAn operational mission could be targeted for 5 cm
• A particular application of PARIS IoD is the direct observation of A particular application of PARIS IoD is the direct observation of tsunamis:tsunamis:
• enhance our knowledge on tsunamis and complement early warning enhance our knowledge on tsunamis and complement early warning systemssystems