Reliability of satellite-to-ground optical transmissions · High to very high data rates ~10 Gbps...
Transcript of Reliability of satellite-to-ground optical transmissions · High to very high data rates ~10 Gbps...
-
Reliability of satellite-to-ground optical
transmissions Lucien Canuet
Jérôme Lacan (ISAE)
Angélique Rissons (ISAE)
Nicolas Védrenne (ONERA)
Géraldine Artaud (CNES)
-
SATELLITE OPTICAL DOWNLINKS - GENERAL CONTEXT
Telecom
Telemetry
& payload
2
Applications of optical links • Data transfer (telemetry/payload) from scientific or defense spacecraft (LEO) • Telecommunication (GEO satellites) • Metropolitan area networks (where fiber optics impractical) • Deep space probes …
-
High to very high data rates
~10 Gbps
>1 Tbps achievable if optical fiber technologies are
exploited (WDM, DWDM, Optical amplifiers etc.) Decongestion of the RF spectrum
Enhanced security (high directivity of the beam)
Stealthy links and jamming capacity reduced
Very large range (« Deep Space » applications)
3
SATELLITE OPTICAL DOWNLINKS - GENERAL CONTEXT
-
High to very high data rates
~10 Gbps
>1 Tbps achievable if optical fiber technologies are
exploited (WDM, DWDM, Optical amplifiers etc.) Decongestion of the RF spectrum
Enhanced security (high directivity of the beam)
Stealthy links and jamming capacity reduced
Very large range (« Deep Space » applications)
Links highly affected by atmospheric turbulence
4
SATELLITE OPTICAL DOWNLINKS - GENERAL CONTEXT
-
High to very high data rates
~10 Gbps
>1 Tbps achievable if optical fiber technologies are
exploited (WDM, DWDM, Optical amplifiers etc.) Decongestion of the RF spectrum
Enhanced security (high directivity of the beam)
Stealthy links and jamming capacity reduced
Very large range (« Deep Space » applications)
Links highly affected by atmospheric turbulence
Improvement of the link budget and telecom performances:
+ Coupling of incident flux into optical fiber (SMF)
+ Adaptive Optics (AO)
+ Optimisation of digital (coding/interleaving) techniques
Joint optimisation of AO and coding techniques to improve reliability of
LEO and GEO downlinks (telemetry and telecoms applications)
5
SATELLITE OPTICAL DOWNLINKS - GENERAL CONTEXT
-
PART I : JOINT OPTIMIZATION – PROBLEM OVERVIEW
PART II : AO FOR OPTICAL DONWLINKS
PART III : PHYSICAL LAYER PERFORMANCE ASSESSMENT
PART IV : AO/CROSS-LAYER OPTIMIZATION
6
PRESENTATION OUTLINE
-
Coupling losses and signal fadings
Atmospheric propagation channel
Scintillation
(Pupil plane) Turbulent Wave front
Plane Wave
7
Phase fluctuations
(Focal plane)
Atmospheric optical turbulence
Δ(Temperature, Pressure,
Humidity) Δ(refractive index) Perturbed wave front
Mitigation techniques ? Coupled flux
Speckles
PSF
Experimental results PhD Thesis Kassem
Saab (CNES/ONERA)
THE ATMOSPHERIC PROPAGATION CHANNEL
Turbulence Fiber Mode
+
-
+Variance résiduelle totale fixe => fixe taux de couplage moyen
+Variation du Drx
+Variation de la proportion des postes d'erreur spatiale et temporelle
8
ADAPTIVE OPTICS PRINCIPLE
SFM
Principle overview
Opto-mechanical system:
Real time correction of
phase distorsions
Three key components:
Deformable mirror
Wavefront sensor
Real-time computer
Perturbed Field
Corrected Field
-
+Variance résiduelle totale fixe => fixe taux de couplage moyen
+Variation du Drx
+Variation de la proportion des postes d'erreur spatiale et temporelle
9
ADAPTIVE OPTICS PRINCIPLE
SFM
Principle overview
Opto-mechanical system:
Real time correction of
phase distorsions
Three key components:
Deformable mirror
Wavefront sensor
Real-time computer
Perturbed Field
Corrected Field
-
10
ADAPTIVE OPTICS PRINCIPLE
SFM
Principle overview
Opto-mechanical system:
Real time correction of
phase distorsions
Three key components:
Deformable mirror
Wavefront sensor
Real-time computer
Perturbed Field
Corrected Field
-
drrr corrres ))²()((12 S
+Variance résiduelle totale fixe => fixe taux de couplage moyen
+Variation du Drx
+Variation de la proportion des postes d'erreur spatiale et temporelle
Minimization of residual phase variance
11
ADAPTIVE OPTICS PRINCIPLE
SFM
Principle overview
Opto-mechanical system:
Real time correction of
phase distorsions
Three key components:
Deformable mirror
Wavefront sensor
Real-time computer
Limitations?
Perturbed Field
Corrected Field
-
12
Time
HIGH PERF. AO
LOW PERF. AO
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES C
ou
ple
d f
lux a
tte
nu
ati
on
-
“BURST”
Raw data stream Code performance limited
13
SYMBOL
50 % Redundancy
Time
HIGH PERF. AO
LOW PERF. AO
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES
CODEWORD
Co
up
led
flu
x a
tte
nu
ati
on
-
“BURST”
Raw data stream
Interleaved data stream
Deinterleaving
Code performance limited
Code performance increased
14
SYMBOL
50 % Redundancy
Time
HIGH PERF. AO
LOW PERF. AO
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES
CODEWORD
Co
up
led
flu
x a
tte
nu
ati
on
-
“BURST”
Raw data stream
Interleaved data stream
Deinterleaving
Code performance limited
Code performance increased
How to effectively allocate memory (interleaver) and redundancy (code)?
15
SYMBOL
50 % Redundancy
CODEWORD
Time
HIGH PERF. AO
LOW PERF. AO
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES C
ou
ple
d f
lux a
tte
nu
ati
on
-
Statistical and temporal characterizations of the channel needed:
instantaneous coupled optical power after partial AO
“BURST”
Raw data stream
Interleaved data stream
Deinterleaving
Code performance limited
Code performance increased
Time
Couple
d f
lux [a.u
.]
16
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES
SYMBOL
HIGH PERF. AO
LOW PERF. AO
50 % Redundancy
Co
up
led
flu
x a
tte
nu
ati
on
CODEWORD
-
Instantaneous mutual information (~channel capacity )
Theoretical maximal data rate at which information can be transferred over the channel given noise level
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES MUTUAL INFORMATION
Bije
ctio
n
The greater the better
17
Mu
tua
l in
form
atio
n
[bits/c
ha
nnel u
se
]
Op
tica
l co
up
led
po
we
r a
tte
nu
ation
-
Series tempo flux im
impact entrelacement-desentrelacement -> capa ergo = moyenne message contr OA pas d’influence sur moy
Motiv mieux que BER cite JL
Appli FSO très peu, jms conj avec OA
Instantaneous mutual information (~channel capacity )
Theoretical maximal data rate at which information can be transferred over the channel given noise level
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES MUTUAL INFORMATION
Emulation of interleaving-deinterleaving at Rx
Sliding average window
Advantages of instantaneous mutual information (MI):
18
Mu
tua
l in
form
atio
n
[bits/c
ha
nnel u
se
]
-
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES MUTUAL INFORMATION
Emulation of interleaving-deinterleaving at Rx
Sliding average window
Ideal infinite interleaver = fundamental limit
Average power losses not recoverable by interleaving
E[MI(t)]
19
Mu
tua
l in
form
atio
n
[bits/c
ha
nnel u
se
] Advantages of instantaneous mutual information (MI):
-
Series tempo flux im
impact entrelacement-desentrelacement -> capa ergo = moyenne message contr OA pas d’influence sur moy
Motiv mieux que BER cite JL
Appli FSO très peu, jms conj avec OA
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES MUTUAL INFORMATION
Emulation of Error Correcting Code (ECC) decoding
Shannon decoding theorem
ECC coding rate = amount of non-redundant
information in message
R0PHY = 0.5
20
Mu
tua
l in
form
atio
n
[bits/c
ha
nnel u
se
] Advantages of instantaneous mutual information (MI):
-
Series tempo flux im
impact entrelacement-desentrelacement -> capa ergo = moyenne message contr OA pas d’influence sur moy
Motiv mieux que BER cite JL
Appli FSO très peu, jms conj avec OA
PHYSICAL LAYER DIGITAL MITIGATION TECHNIQUES MUTUAL INFORMATION
Emulation of Error Correcting Code (ECC) decoding
Shannon decoding theorem
Not Decoded = Link Outage
Successfully Decoded
PHY layer performance metric = Outage probability
21
Mu
tua
l in
form
atio
n
[bits/c
ha
nnel u
se
] Advantages of instantaneous mutual information (MI):
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH
22
CODING possible at both PHY LAYER and HIGHER LAYERS
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
APPLICATION Rx
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH
23
CODING possible at both PHY LAYER and HIGHER LAYERS
Erasure Code
(packets)
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
Error Corr. Code
(bits)
INFO Redundancy INFO Redundancy Red. Redundancy
INFO INFO Red.
Red.
Redundancy
HL Erasure code redundancy
PHY Error correcting code redundancy
APPLICATION Rx
+ … …
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH
24
Erasure Code
&
Packet interleaving
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
Error Corr. Code
&
Symbol interleaving
APPLICATION Rx
Perf. metric: PER
Perf. metric: MI/Outage Prob.
CODING and INTERLEAVING possible at both PHY LAYER and HIGHER LAYERS
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH
25
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
APPLICATION Rx
Perf. metric: PER
Perf. metric: MI/Outage Prob.
CODING and INTERLEAVING possible at both PHY LAYER and HIGHER LAYERS
Benefits of (optimized) cross-layer coding scheme:
• Lowering Packet Error Rate (PER)
• Alleviate drawbacks required by long interleavers needed on bursty channel
Erasure Code
&
Packet interleaving
Error Corr. Code
&
Symbol interleaving
-
Developed module
Enhanced/Validated module
End-to-end simulation modeling
Simplified simulation modeling
Analytic modeling
Adaptive
Optics
Atmos.
Propagation
Channel
Higher
Layers
Tx
Application
Tx Data
Application
Rx Data
Higher
Layers
Rx
Modulator Demodulator Channel
encoder
Channel
decoder
Physical Layer
OVERALL SYSTEM OVERVIEW
26
-
PART I : JOINT OPTIMIZATION – PROBLEM OVERVIEW
PART II : AO FOR OPTICAL DONWLINKS
Partial AO Analytic Modeling
Analytic Modeling of Partially Corrected Coupled Flux into SMFs
PART III : PHYSICAL LAYER PERFORMANCE ASSESSMENT
PART IV : AO/CROSS-LAYER OPTIMIZATION
27
Adaptive
Optics
Atmos.
Propagation
Channel
PRESENTATION OUTLINE
-
Analytical laws for each one of the errors are known
2ph & ron
+ 2Alias + 2scintill.
2fitting Fitting Error
+ 2PS
+ 2Tempo Temporal Error
drrr corrres ))²()((12 S
+Variance résiduelle totale fixe => fixe taux de couplage moyen
+Variation du Drx
+Variation de la proportion des postes d'erreur spatiale et temporelle
Minimization of residual phase variance
28
PARTIAL ADAPTIVE OPTICS
Intrinsic limitations = partial AO
Aliasing Error
SFM Neglect impact of noise and scintillation on phase sensor
error
Perturbed Field
Corrected Field
-
Uncorrected high
frequencies Perfectly corrected
low frequencies
Spatial Frequencies
Pow
er S
pec
tra
l D
ensi
ty
Fitting
Error
Ideal AO correction system: the low order residuals are perfectly corrected
Total residual error
29
PARTIAL ADAPTIVE OPTICS
-
Uncorrected high
frequencies
Patially corrected
low frequencies
Spatial Frequencies
Pow
er S
pec
tra
l D
ensi
ty
Fitting
Error
Aliasing+Temporal
Errors
Partial AO correction system: low order residuals that are highly correlating
Total residual error Non negligible impact
on communication
performances
30
PARTIAL ADAPTIVE OPTICS
-
31
INSTANTANEOUS COUPLED OPTICAL POWER ATTENUATION
Neglecting amplitude spatial structures influence on coupling fluctuations:
Scintillation Injection losses Average
coupling losses
Collected power
fluctuations Rough approx, validated for medium elevation, small perturbations
-
32
Optical power attenuation probability density distribution Average fading time against normalized threshold
Canuet L., Védrenne N., Conan J-M. , Petit C., Artaud G., Rissons A., and Lacan J., « Statistical properties of single-mode fiber
coupling of satellite-to-ground laser links partially corrected by adaptive optics » J. Opt. Soc. Am. A 35, 148-162 (2018)
INSTANTANEOUS COUPLED OPTICAL POWER ATTENUATION
Neglecting amplitude spatial structures influence on coupling fluctuations:
Scintillation Injection losses Average
coupling losses
Collected power
fluctuations Rough approx, validated for medium elevation, small perturbations
Illustration GEO-Downlink, Drx = 50cm, r0= 0.069 m, 9 radial orders at 1 kHz
-
PART I : JOINT OPTIMIZATION – PROBLEM OVERVIEW
PART II : AO FOR OPTICAL DONWLINKS
PART III : PHYSICAL LAYER PERFORMANCE ASSESSMENT
Optical Communication Subsystem Overview
Outage Probability
Minimum Required Interleaver
PART IV : AO/CROSS-LAYER OPTIMIZATION
33
Adaptive
Optics
Atmos.
Propagation
Channel
Modulator Demodulator Channel
encoder
Channel
decoder
Physical Layer
PRESENTATION OUTLINE
-
PHYSICAL LAYER PERFORMANCE END-TO-END MODELING
34
Adaptive
Optics
Atmos.
Propagation
Channel
Modulator Demodulator Channel
encoder
Channel
decoder
Physical Layer
EDFA
PREAMPLIFIER
OPTICAL
FILTER
EDFA
PREAMPLIFIER
OPTICAL
FILTER
PIN
PHOTODIODE
ELECTRICAL
FILTER
OOK Preamplified Detector
BALANCED
DETECTOR
ELECTRICAL
FILTER
DELAY-LINE
INTERFEROMETER
DPSK Balanced Detector
TB
-
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m
35
Target:
Min required power
Outage probability for 10 Gbps link coderate R0=0.5
Med. Perf. AO
15 actuators 800 Hz
5dB avg. attenuation
Low. Perf. AO
10 actuators 500 Hz
7dB avg. attenuation
DPSK system with -7dB AO
DPSK system with -5dB AO
-
36
Target:
Outage probability for 10 Gbps link coderate R0=0.5
Med. Perf. AO
15 actuators 800 Hz
5dB avg. attenuation
Low. Perf. AO
10 actuators 500 Hz
7dB avg. attenuation
DPSK system with -5dB AO
DPSK system with -7dB AO
30 ms interleaver (300 Mb)
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m
-
37
Target:
Outage probability for 10 Gbps link coderate R0=0.5
30 ms interleaver (300 Mb)
3 dB
Optimizing interleaver at targeted outage probability & required power?
Med. Perf. AO
15 actuators 800 Hz
5dB avg. attenuation
Low. Perf. AO
10 actuators 500 Hz
7dB avg. attenuation
OOK system with -3db AO
DPSK system with -5dB AO
DPSK system with -7dB AO
High Perf. AO
50 actuators at 2 kHz
3dB avg. attenuation
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m
-
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m | OOK
Code rate
38
Minimum Required PHY interleaving depth High Perfo. AO | Outage Prob. 10-2
PHY Code Rate
Min
. R
equir
ed P
HY
Inte
rlea
ver
Mem
ory
[M
b]
Low redundancy
Req. Rx Power = -41 dBm
Req. Rx Power = -38 dBm
-
Code rate
39
Min
. R
equir
ed P
HY
Inte
rlea
ver
Mem
ory
[M
b]
Req. Rx Power = -41 dBm
Req. Rx Power = -38 dBm
Minimum Required PHY interleaving depth High Perfo. AO | Outage Prob. 10-2
Partial AO
Ideal AO
Neglecting lower order residuals underestimate required memory by 50%
PHY Code Rate
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m | OOK
-
Code rate
40
PHY Code Rate
Minimum Required PHY interleaving depth High Perfo. AO | Outage Prob. 10-2
Ideal Infinite interleaver limit
Min
. R
equir
ed P
HY
Inte
rlea
ver
Mem
ory
[M
b]
Req. Rx Power = -41 dBm
Req. Rx Power = -38 dBm
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m | OOK
-
Code rate
41
PHY Code Rate
Minimum Required PHY interleaving depth High Perfo. AO | Outage Prob. 10-2
Decrease in redundancy
sharp increase in memory
Min
. R
equir
ed P
HY
Inte
rlea
ver
Mem
ory
[M
b]
Req. Rx Power = -41 dBm
Req. Rx Power = -38 dBm
Transfer to HL
PHYSICAL LAYER TRADE-OFF ASSESSMENT
LEO application case Drx = 0.25 m | r0 = 0.056 m | OOK
-
PART I : JOINT OPTIMIZATION – PROBLEM OVERVIEW
PART II : AO FOR OPTICAL DONWLINKS
PART III : PHYSICAL LAYER PERFORMANCE ASSESSMENT
PART IV : AO/CROSS-LAYER OPTIMIZATION
Modeling of Cross-layer Coding Scheme
Case Study: LEO Downlink Using DPSK
42
Adaptive
Optics
Atmos.
Propagation
Channel
Higher
Layers
Tx
Higher
Layers
Rx
Modulator Demodulator Channel
encoder
Channel
decoder
Physical Layer
PRESENTATION OUTLINE
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH : Overview
43
CODING and INTERLEAVING possible at both PHY LAYER and HIGHER LAYERS
Erasure code
R0HL
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
Error correcting code
R0PHY
Red.
Redundancy
HL Redundancy
PHY
Redundancy APPLICATION Rx
Perf. metric: MI/Outage Prob.
Perf. metric: PER
Fixed global code rate
R0GLOBAL = R
0PHYR
0HL
Useful Data Rate = raw data-rate x R0GLOBAL
The PHY and HL coderates inherently related through definition of global coderate:
-
INFO Redundancy
INFO Redundancy
Error corr. code
APPLICATION PHY LAYER HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
CROSS-LAYER APPROACH : Overview
44
Red.
Redundancy
HL Redundancy
PHY
Redundancy
Fixed global code rate
R0GLOBAL = R
0PHYR
0HL
Useful Data Rate = raw data-rate x R0GLOBAL
The PHY and HL coderates inherently related through definition of global coderate:
Tx HL memory = HL interleaver depth x raw data-rate x R0PHY
Tx PHY memory = PHY interleaver depth x raw data-rate Advantage to HL
CODING and INTERLEAVING possible at both PHY LAYER and HIGHER LAYERS
Erasure code
R0HL
APPLICATION Tx
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
TRANSPORT
NETWORK
DATA LINK
PHYSICAL
Higher layers
FSO
Channel
Error correcting code
R0PHY
APPLICATION Rx
Perf. metric: MI/Outage Prob.
Perf. metric: PER
-
PHYSICAL LAYER
R0PHY = 0.5
CROSS-LAYER APPROACH : Modeling R
0GLOBAL = R
0PHYR
0HL = 0.3
45
Mu
tua
l in
form
atio
n
-
PHYSICAL LAYER
HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
R0PHY = 0.5
46
Mu
tua
l in
form
atio
n
PH
Y o
utp
ut P
ER
CROSS-LAYER APPROACH : Modeling R
0GLOBAL = R
0PHYR
0HL = 0.3
-
PHYSICAL LAYER
HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
R0PHY = 0.5
47
BAD
GOOD
Mu
tua
l in
form
atio
n
PH
Y o
utp
ut P
ER
CROSS-LAYER APPROACH : Modeling R
0GLOBAL = R
0PHYR
0HL = 0.3
1-R0HL = 0.4
-
PHYSICAL LAYER
HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
APPLICATION LAYER
R0PHY = 0.5
1-R0HL = 0.4
48
Mu
tua
l in
form
atio
n
PH
Y o
utp
ut P
ER
H
L o
utp
ut
Decoded P
ER
CROSS-LAYER APPROACH : Modeling R
0GLOBAL = R
0PHYR
0HL = 0.3
-
PHYSICAL LAYER
HIGHER LAYERS DATA LINK | NETWORK |
TRANSPORT
APPLICATION LAYER FINAL PERFORMANCE METRIC : Average of decoded PER
R0PHY = 0.5
1-R0HL = 0.4
49
Mu
tua
l in
form
atio
n
PH
Y o
utp
ut P
ER
H
L o
utp
ut
Decoded P
ER
CROSS-LAYER APPROACH : Modeling R
0GLOBAL = R
0PHYR
0HL = 0.3
-
Average decoded PER against size of HL interleaving
50
Tx Higher Layers Interleaving Memory [Mb]
CROSS-LAYER APPROACH : Benefits illustration
R0
GLOBAL = 0.8 | Req. Power = -38 dBm | PHY interleaver 50 Mb
R0PHY = 0.8 R
0HL = 1
R0PHY = 0.82 R
0HL = 0.97
R0PHY = 0.86 R
0HL = 0.93
R0PHY = 0.90 R
0HL = 0.89
-
Average decoded PER against size of HL interleaving
51
Tx Higher Layers Interleaving Memory [Mb]
CROSS-LAYER APPROACH : Benefits illustration
R0
GLOBAL = 0.8 | Req. Power = -38 dBm | PHY interleaver 50 Mb
R0PHY = 0.8 R
0HL = 1
R0PHY = 0.82 R
0HL = 0.97
R0PHY = 0.86 R
0HL = 0.93
R0PHY = 0.90 R
0HL = 0.89
No coding on HL
-
52
Tx Higher Layers Interleaving Memory [Mb]
CROSS-LAYER APPROACH : Benefits illustration
R0
GLOBAL = 0.8 | Req. Power = -38 dBm | PHY interleaver 50 Mb
R0PHY = 0.8 R
0HL = 1
R0PHY = 0.82 R
0HL = 0.97
R0PHY = 0.86 R
0HL = 0.93
R0PHY = 0.90 R
0HL = 0.89
Optimal allocation of
redundancy btw PHY and HL
Average decoded PER against size of HL interleaving
-
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4 | 10 Gbps LEO link
DPSK Receiver
53
1000
800
600
400
200
0
Min
. To
tal T
x M
em
ory
(P
HY
+H
L)
[Mb
]
2 3 4 5 6 7 8 9 No coding
(Redundancy)
A lot of coding
(Redundancy) Useful Data Rate [Gbps]
High Perf. AO (-3 dB)
Med Perf. AO (-5 dB)
Low Perf. AO (-7 dB)
-
54
High Perf. AO (-3 dB)
Med Perf. AO (-5 dB)
Low Perf. AO (-7 dB)
1000
800
600
400
200
0
Min
. To
tal T
x M
em
ory
(P
HY
+H
L)
[Mb
]
2 3 4 5 6 7 8 9 No coding
(Redundancy)
A lot of coding
(Redundancy) Useful Data Rate [Gbps]
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4 | 10 Gbps LEO link
DPSK Receiver
-
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 97
PHY Coderate : 0.48 | HL Coderate : 0.62
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 37
PHY Coderate : 0.36 | HL Coderate : 0.83
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 87
PHY Coderate : 0.85 | HL Coderate : 0.94
HL interl. exclusively
55
1000
800
600
400
200
0
Min
. To
tal T
x M
em
ory
(P
HY
+H
L)
[Mb
]
2 3 4 5 6 7 8 9 No coding
(Redundancy)
A lot of coding
(Redundancy) Useful Data Rate [Gbps]
High Perf. AO (-3 dB)
Med Perf. AO (-5 dB)
Low Perf. AO (-7 dB)
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4 | 10 Gbps LEO link
DPSK Receiver
Unfad. Rx Power [dBm] : -41
PHY mem. [Mb] : 102 | HL mem. [Mb] : 0
PHY Coderate : 0.5 | HL Coderate : 1
-
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 97
PHY Coderate : 0.48 | HL Coderate : 0.62
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 37
PHY Coderate : 0.36 | HL Coderate : 0.83
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 87
PHY Coderate : 0.85 | HL Coderate : 0.94
HL interl. exclusively
56
1000
800
600
400
200
0
Min
. To
tal T
x M
em
ory
(P
HY
+H
L)
[Mb
]
2 3 4 5 6 7 8 9 No coding
(Redundancy)
A lot of coding
(Redundancy) Useful Data Rate [Gbps]
High Perf. AO (-3 dB)
Med Perf. AO (-5 dB)
Low Perf. AO (-7 dB)
HL interleaving optimal in less challenging conditions
Physical turbulence mitigation techniques (AO) can have significant impact on overall system
design optimization by driving hardware trade-offs (eg. ASIC vs RAM interleaver)
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4 | 10 Gbps LEO link
DPSK Receiver
Unfad. Rx Power [dBm] : -41
PHY mem. [Mb] : 102 | HL mem. [Mb] : 0
PHY Coderate : 0.5 | HL Coderate : 1
-
Developed module
Enhanced/Validated module
End-to-end simulation modeling
Simplified simulation modeling
Analytic modeling
Adaptive
Optics
Atmos.
Propagation
Channel
Higher
Layers
Tx
Application
Tx Data
Application
Rx Data
Higher
Layers
Rx
Modulator Demodulator Channel
encoder
Channel
decoder
Physical Layer
SIGNIFICANT RESULTS
57
• First accurate model of partially corrected coupled flux into SMFs
• Detailed investigation of required physical layer interleaving depth and ECC
• First application of cross-layer coding/interleaving scheme to sat. opt. transmissions
• Investigation of overall optimization of AO and data reliability mechanisms
CONCLUSION
How to ensure reliable optical downlinks?
-
RECOMMENDATION FOR FUTURE WORK
Experimental validation- LEO Downlink planned (DLR’s OSIRIS & ONERA’s LISA)
Investigation of performance evolution over the duration of a whole link
Input turbulence conditions not well known (except astron. observation
sites)
Transposition of approach to GEO uplink
58
-
PUBLICATIONS AND COMMUNICATIONS
59
Peer-reviewed publication Canuet L., Védrenne N., Conan J-M. , Petit C., Artaud G., Rissons A., and Lacan J., « Statistical properties of
single-mode fiber coupling of satellite-to-ground laser links partially corrected by adaptive optics » J. Opt.
Soc. Am. A 35, 148-162 (2018)
International conferences Canuet L., Védrenne N., Conan J-M., Artaud G., Rissons A., and Lacan J., “Evaluation of communication
performance for adaptive optics corrected GEO-to-ground laser links”, in proceedings of ICSO (International
Conference on Space Optics) 2016, Biarritz, France
Canuet L., Lacan J., Védrenne N., Artaud G., and Rissons A., “Performance Evaluation of Coded Transmission
for Adaptive-Optics Corrected Satellite-To-Ground Laser Links”, in proceedings of ICSOS (IEEE International
Conference on Space Optical Systems and Applications), November 2017 – Naha, Okinawa, Japan
Canuet L., Lacan J., Védrenne N., Artaud G., and Rissons A., “Cross-layer optimization for adaptive-optics
corrected satellite-to-ground laser links”, in proceedings of the 8th international symposium – OPTRO 2018
optronics in defense and security, 6-8 February 2018, Paris
-
ACKNOWLEDGEMENTS
Committee members Pr. Peucheret
Pr. Belmonte
Pr. Kahn
Dr. Sodnik
Dr. Perlot
Advisors Jérôme Lacan
Angélique Rissons
Nicolas Védrenne
Géraldine Artaud
ONERA Jean-Marc Conan
Cyril Petit
CNES Bouchra Benammar
Airbus DS Sylvain Poulenard
Hervé Haag
Thomas Anfray
Thales AS
Michel Sotom
Arnaud Le Kernec
TéSA Corinne Mailhes
-
PHYSICAL LAYER TRADE-OFF ASSESSMENT Minimum required interleaving depth
Code rate
61
PHY Code Rate
Min
. R
equir
ed P
HY
Inte
rlea
ver
Mem
ory
[M
b]
Req. Rx Power = -41 dBm
Req. Rx Power = -38 dBm
Minimum Required PHY interleaving depth High Perfo. OA | Outage Prob. 10-2
Analytic estimation
Analytic estimation based on statistical and temporal characteristics of coupled flux
-
0.60 | 1.00 0.65 | 0.92 0.70 |
0.92
0.75 | 0.80 0.80 |
0.75
0.85 | 0.70 0.90 | 0.67 0.95 | 0.62 1.00 | 0.60
PHY coderate | HL coderate
Min
. T
X T
ota
l M
em
ory
[M
b]
1000
0
RX Total mem. [Mb] :
102
PHY mem. [Mb] : 102
HL mem. [Mb] : 0
NOT ERROR-FREE FOR
CONSIDERED INTERLEAVERS
RANGE
MINIMUM MEMORY REQUIRED FOR ERROR-FREE TRANSMISSION
10 Gbps link | R0
GLOBAL = 0.6 (6 Gbps throughput) | High Perf. AO
62
RX Total mem. [Mb] :
678
PHY mem. [Mb] : 502
HL mem. [Mb] : 176 Functioning point
requiring both PHY and
HL interleaving
-
0.60 | 1.00 0.65 | 0.92 0.70 | 0.92 0.75 | 0.80 0.80 | 0.75 0.85 | 0.70 0.90 | 0.67 0.95 | 0.62 1.00 | 0.60
NOT ERROR-FREE FOR
CONSIDERED INTERLEAVERS
RANGE
Region of HL Interleaving only
RX Total mem. [Mb] : 88
PHY mem. [Mb] : 0
HL mem. [Mb] : 88 RX Total mem. [Mb] : 65
PHY mem. [Mb] : 0
HL mem. [Mb] : 65
Optimal allocation of
redundancy
PHY coderate | HL coderate
Min
. T
X T
ota
l M
em
ory
[M
b]
1000
0
Functioning point requiring
minimum overall memory and
power
RX Total mem. [Mb] : 972
PHY mem. [Mb] : 0
HL mem. [Mb] : 972
MINIMUM MEMORY REQUIRED FOR ERROR-FREE TRANSMISSION
10 Gbps link | R0
GLOBAL = 0.6 (6 Gbps throughput) | High Perf. AO
63
-
Unfad. Rx Power [dBm] : -40
PHY mem. [Mb] : 202 | HL mem. [Mb] : 0
PHY Coderate : 0.30 | HL Coderate : 1
Total Latency [ms] : 20.25
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 40
PHY Coderate : 0.39 | HL Coderate : 0.77
Total Latency [ms] : 10.25
Unfad. Rx Power [dBm] : -39
PHY mem. [Mb] : 102 | HL mem. [Mb] : 0
PHY Coderate : 0.5 | HL Coderate : 1
Total Latency [ms] : 10.25
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 0
PHY Coderate : 0.5 | HL Coderate : 1
Total Latency [ms] : 0
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 202 | HL mem. [Mb] :
0
PHY Coderate : 0.7 | HL Coderate : 1
Total Latency [ms] : 20.25
Unfad. Rx Power [dBm] : -41
PHY mem. [Mb] : 102 | HL mem. [Mb] : 0
PHY Coderate : 0.7 | HL Coderate : 1
Total Latency [ms] : 10.25
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 102 | HL mem. [Mb] :
0
PHY Coderate : 0.9 | HL Coderate : 1
Total Latency [ms] : 10.25
HL intel.
exclusively
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4
OOK Receiver
-
Increase in Drx
=
Aperture averaging of scintillation
Adaptive Optics
=
Real-time correction of phase fluctuations only
+
For a fixed Drx scintillation effects are imposed upon the detector
The design of the AO system must be done
consequently
Trade-offs scintillation/phase fluctuations mitigation and therefore performances/cost
High perfs AO Limited by scintillation Low perfs AO Limited by phase fluctuations
65
PHYSICAL MITIGATION TECHNIQUES : APERTURE AVERAGING PLUS
AO
Cou
pled
flux
Cou
pling
loss
es
Irradi
ance
Atten
uatio
n
[a.u
.]
Ti
m
e
[s]
Coupled flux Coupling losses Irradiance
Time [s]
Atten
uation
[a
.u.]
Time [s]
Atten
uation
[a
.u.]
-
EM field of incoming wave
Unperturbed amplitude Log-amplitude
Phase (corrected)
Instantaneous coupled optical power attenuation
Neglecting amplitude spatial structures influence on coupling fluctuations:
Scintillation Injection losses
Pupil transmittance SMF mode
Rustic approx, justified for GEO downlinks (medium elevation, small perturbations)
Cou
ple
d flu
x [a
.u.]
Time [s]
End-to-end wave optics simulation results
66
INSTANTANEOUS COUPLED OPTICAL POWER ATTENUATION
Simplified Model
Wave Optics σ2 = 0.001
Perfect AO correction
|Difference|
Average
coupling losses
Collected power
fluctuations
-
Phase decomposition into Zernike polynomials (= Orthonormal basis over a plane and circular domain)
Zernike coefficient #i Zernike polynomial #i
Phase decomposition after “re-orthonormalisation”
Retro-propagated SMF mode in pupil plane
Set of Zernike polynomials not orthonormal
Statistical properties (PDF) of each ai : known analytically (Independent Gaussian variables BUT not identical)
Temporal properties (PSD) of each ai : known analytically
Transfer Matrix {ai}{bi} : known analytically
“Spatial variance” of the phase fluctuations
in the focal plane
Closed-form injection efficiency approximation :
Injection efficiency
without aberrations
67
INJECTION LOSSES STATISTICAL CHARACTERIZATION
-
Scintillation Injection losses
Instantaneous coupled optical power attenuation
+ Temporal properties (PSD, coherence time)
Analytic expressions
Statistical laws (PDF/CDF)
Typical Log-normal distribution “Exponentiated sum of
non-identical Gamma variates”
68
INJECTION LOSSES STATISTICAL CHARACTERIZATION
-
Unfad. Rx Power [dBm] : -38
PHY mem. [Mb] : 0 | HL mem. [Mb] : 37
PHY Coderate : 0.36 | HL Coderate : 0.83
69
High Perf. AO (-3 dB)
Med Perf. AO (-5 dB)
Low Perf. AO (-7 dB)
1000
800
600
400
200
0
Min
. To
tal T
x M
em
ory
(P
HY
+H
L)
[Mb
]
2 3 4 5 6 7 8 9 No coding
(Redundancy)
A lot of coding
(Redundancy) Useful Data Rate [Gbps]
Unfad. Rx Power [dBm] : -41
PHY mem. [Mb] : 102 | HL mem. [Mb] : 0
PHY Coderate : 0.5 | HL Coderate : 1
Unfad. Rx Power [dBm] : -39
PHY mem. [Mb] : 102 | HL mem. [Mb] :0
PHY Coderate : 0.7 | HL Coderate : 1
RESOURCES OPTIMIZATION USING MINIMUM TOTAL MEMORY
TARGET PER = 10-4 | 10 Gbps LEO link
DPSK Receiver