• Aucun résultat trouvé

Performance Analysis of Fractionally Spaced Equalization in Non-linear Multicarrier Satellite Channels

N/A
N/A
Protected

Academic year: 2021

Partager "Performance Analysis of Fractionally Spaced Equalization in Non-linear Multicarrier Satellite Channels"

Copied!
24
0
0

Texte intégral

(1)

R. Piazza , M. R. Bhavani Shankar ,T. Berheide, M. Grasslin, and S. Cioni

32th AIAA International Communications Satellite System Conference (ICSSC), San Diego

August 5th, 2014

Performance Analysis of Fractionally Spaced

Equalization in Non-linear Multicarrier Satellite

Channels

(2)

Presentation Overview

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(3)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(4)

Scenario

• Multicarrier gateway uplink

• Multicarrier transponder

– Joint input /output filtering – Joint power amplification

• Advantages:

– HW saving – Weight saving – Flexibility

(5)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(6)

Channel Impairments and System Constraints

• Power & Spectral Efficiency Trade off

• System Contraints:

– No On-board Signal Processing – Low complexity User Terminals

INTERMODULATION PRODUCTS

(7)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(8)

• Countermeasures Techniques:

1. Multicarrier Predistortion at the gateway 1 2. Advanced Receiver Processing:

1. Fractionally Spaced Equalization

2. Optimized de-mapping at the user terminals (UT)

1R. Piazza, E. Zenteno, and e. al, “Multicarrier digital predistortion/equalization techniques for non-linear satellite channels,”in Proc. 30th AIAA Intern.Commun. Satellite Syst. Conference (ICSSC), Ottawa, Canada, Sep. 2012.

System Architecture

HPA p1

pm

pM

p1

pm

pM

f1

fm

fM

f1

fm

fM u1

um

uM

y1

ym

yM

η

DPD

. . .. . .

. . .. . .

. . .. . .

UPLINK DOWNLINK

RX1

RXm

RXM GATEWAY

IMUX OMUX

SATELLITE TRANSPONDER

x1

xm

xM

EQ

EQ

EQ

r1

rm

rM

(9)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(10)

Fractionally Spaced Equalization

• Receiver equalization that exploits higher

sampling rate

2

:

– K>1 samples per symbol

• It aims to compensate for:

– Non-constant group delay of the channel – Residual Linear and non-linear distortions – Non-optimal receiver sampling

2R. D. Gitlin, S. B. Weinstein, “Fractionally Spaced Equalization: An Improved Digital Transversal Equalizer,” Bell Systems Technical Journal, 60:2, February 1981, available online http://archive.org/details/bstj60-2-275.

(11)

FSE Architecture

• FSE as Linear Filtering:

– rm n = bk1 m k1 𝑣𝑚 n − k1 = 𝒃𝑚𝒗𝑚(𝑛)

• Parameters Estimation:

– 𝒃𝑚 = argminbm 𝑁𝑛=1 𝐸 𝑟𝑚 𝑛 − 𝑢𝑚 𝑛 2 – Standard Least Squares Solution

– Adaptive and based on pilots

(12)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(13)

Non-linear Bias in the RX symbols

• Equalized symbols shows some residual non-linear bias w.r.t. to the reference constellation

• This bias degrades the bit error rate (BER) performance

– Need to determine more accurate reference constellation for decoding

-1.5 -1 -0.5 0 0.5 1 1.5

-1.5 -1 -0.5 0 0.5 1 1.5

I

Q

IBO=4

Noiseless Scatter Plot DVB-S2 Constellation Estimated Centroids

(14)

Average and Centroids based De-mapping

• For linear systems average constellation de-

mapping (ACD) suffices:

– One scaling factor : 𝛽 = 𝑎𝑟𝑔min

𝑐

𝑥−𝑐 2

𝑀 𝑥∈ℱ𝑘 𝑘=1

𝑎𝑘 2

𝑀𝑘=1

• For a general non-linear system we need one-

to one re-mapping (CBD):

– For each constellation point 𝑘:

Centroid 𝑐𝑘 = 𝑎𝑟𝑔min

𝑐 𝑥∈ℱ𝑘 𝑥 − 𝑐 2 , 𝑘 ∈ [1, 𝑀]

– Centroids estimation based on pilots

(15)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(16)

TD = E

b

N

0 NL

− E

b

N

0 Ideal

+ OBO.

• Total Degradation:

– Evaluated at a Target Packet Error Rate

– Spectral Efficiency & HPA Power efficiency

where

Figure Of Merit

HPA efficiency loss

Code loss OBO = P

out

/P

sat

𝑑𝐵

(17)

TD Performance Two Carriers

Satellite Channel (1)

• Setting: 16.36 Mbaud, 16 APSK, Roll-off=0.2,

LDPC with Code Rate=3/4, Transp. BW=36MHz

1.0 1.5 2.0 2.5 3.0

3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3

Output Back Off [dB]

Total Degradation [dB]

MC_DPD + EQ MC_DPD + FSE_ACD MC_DPD + FSE_CBD

• EQ: Baseline symbols spaced equalization

• FSE with Average Const.

Decoding provides 0.2 dB over EQ

• Centroids decoding

provides additional 0.15 dB

• Total of 0.3-0.4 dB of gain

(18)

TD Performance Two Carriers

Satellite Channel (2)

• Setting: 18 Mbaud, 16 APSK, Roll-off=0.2,

LDPC with Code Rate=3/4, Transp. BW=36MHz

1.0 1.5 2.0 2.5 3.0 3.5

3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0

Output Back Off [dB]

Total Degradation [dB]

MC_DPD + EQ MC_DPD + FSE_ACD MC_DPD + FSE_CBD

• Higher baud-rate leads to higher degradation

• FSE with Average Const.

Decoding provides 0.15dB over EQ

• Centroids decoding

provides additional 0.25 dB

• Total of ~0.4 dB of gain

(19)

TD Performance Triple Carriers

Satellite Channel (1)

• Setting: 10 Mbaud, 16 APSK, Roll-off=0.2,

LDPC with Code Rate=3/4,Transp. BW=36MHz

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8

3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8

Output Back Off [dB]

Total Degradation [dB]

MC_DPD + EQ: EXT MC_DPD + EQ: INT MC_DPD + FSE_ACD: EXT MC_DPD + FSE_ACD: INT MC_DPD + FSE_CBD: EXT MC_DPD + FSE_CBD: INT

• External carrier degraded by the MUX filters edge

• FSE with Average Const.

Decoding provides up to 0.1dB over EQ

• Centroids decoding

provides additional~ 0.15 dB

• Total ~0.2 dB of gain

(20)

TD Performance Triple Carriers

Satellite Channel (2)

• Setting: 10 Mbaud, 32 APSK, Roll-off=0.2,

LDPC with Code Rate=4/5,Transp. BW=36MHz

1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

4.0 4.5 5.0 5.5 6.0

Output Back Off [dB]

Total Degradation [dB]

MC_DPD + EQ: EXT MC_DPD + EQ: INT MC_DPD + FSE_ACD:EXT MC_DPD + FSE_ACD:INT MC_DPD + FSE_CBD: EXT MC_DPD + FSE_CBD: INT

• Higher Spectral efficiency leads to higher

degradation

• FSE with Average Const.

Decoding provides up to 0.25 dB to over EQ

• Centroids decoding

provides additional~ 0.25 dB

• Total ~0.5 dB of gain

(21)

Robustness to Sampling Error:

Central carrier of a Three Carrier Channel

• Setting: 16 APSK, Roll Off=0.2, IBO=4 dB, LDPC

with Code Rate=3/4,Transp. BW=36MHz

12 12.5 13 13.5 14 14.5 15 15.5 16 16.5 17

10-5 10-4 10-3 10-2 10-1

Es/No[dB]

BER

22% Sampling Error:MC DPD + EQ 22% Sampling Error:MC DPD + FSE ACD Perfect Timing: MC-DPD EQ12Ts

35% Sampling Error: MC DPD + FSE ACD 35% Sampling Error: MC DPD+EQ

• Standard EQ is very sensitive to sampling

• FSE compensates substantially for the sampling error

– 22% :perfect recovery – 35%: only 0.5 dB of loss

EQ SNR LOSS FSE SNR LOSS

(22)

Outline

1. Scenario

2. Channel impairments and System Constraints 3. System Architecture

4. Fractionally Spaced Equalization 5. Optimized Receiver De-mapping 6. Simulations

7. Conclusions

8. Acknowledgments

(23)

Conclusion

• UT FSE equalization evaluated for multicarrier

satellite channels:

– Provides about ~0.1/0.2 dB of gain when

multicarrier predistortion is applied at the GW – Is shown to be robust with respect to sampling

accuracy

• Optimized Symbols de-mapping:

– Provides additional ~0.1-0.3 dB of gain – Low complexity

(24)

Acknowledgement & Contact

The authors would like to thank the European Space Agency (ESA) for their support through the ARTES 5.1 project “On Ground Multicarrier Digital Equalization/Predistortion Techniques for Single or Multi Gateway

Applications” (APEXX) – ESA Contract No.: 4000105192/12/NL/AD.

Contact Details:

• Roberto Piazza: roberto.piazza@uni.lu

Références

Documents relatifs

The non-linear channel impairments, resulting in a distortion of the transmitted signals, have been tackled in the literature in different ways, based on peak-to-average

In this work, we study the performance variation of a multicarrier non-linear satellite channel to different channel parameters with DPD and Equalization in place to

A typical satellite resource sharing scenario is the joint amplification of multiple channels/ carriers using a single wideband HPA (High Power Amplifier) onboard a

3 plots the BER performance for four turbo-iterations for both the time and frequency domain MMSE equalizers: the exact MMSE linear equalizer, the time domain MMSE equalizer

Multicarrier data predistortion is considered at the transmitting gateway in combination with single carrier symbol rate equalization at the receiving user terminals 12.

The rest of the paper is organized as follows: Section 2 defines the non-linear channel model, Section 3 describes the novel predistortion technique, Section 4 provides a tai-

SRS DPD uses a multi-rate processing algorithm to compute basis functions that enhances non-linear mitigation features of data DPD, while it reduces the digital processing

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of