Radio overfiber tutorial_iwt_2013_nggo

1,768 views
1,598 views

Published on

Published in: Business, Technology
0 Comments
6 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,768
On SlideShare
0
From Embeds
0
Number of Embeds
7
Actions
Shares
0
Downloads
0
Comments
0
Likes
6
Embeds 0
No embeds

No notes for slide
  • In the central office (CO), two different real valued 64-subcarrier 4-QAM OFDM baseband signals with 198.5 Mb/s net data rate (excluding the training symbols) and 312 MHz of bandwidth are generated by an arbitrary waveform generator (ArbWaveGen). The OFDM symbols are arranged in frames of 10 symbols. The first 3 symbols implement the trainingsequence, and 10% cyclicprefix is added. A dualchannel baseband MIMO-OFDM signal is generated in the ArbWaveGen, which is thenup-converted to a 5.65 GHz radio frequency (RF) carrier; the signal in one arm is up-convertedusing a mixer and the other arm implements RF up-conversionusing the vector signal generator (VSG).
  • Radio overfiber tutorial_iwt_2013_nggo

    1. 1. Radio-over-fiber Idelfonso Tafur Monroy E-mail: idtm@fotonik.dtu.dk Neil Guerrero Gonzalez CPqD
    2. 2. 2 Metro-access and short range communications group
    3. 3. 3 Team members Staff (6) Idelfonso Tafur Monroy, Prof. Darko Zibar, Assoc. Prof. Jesper B. Jensen, Asst. Prof. J.J.Vegas Olmos, Asst. Prof. Antonio Caballero, Postdoc PhD Researchers (12 ) Cyd Delgado Jose Estaran Bomin Li Valeria Arlunno Xiaodan Pang Alexander Lebedev Maisara Othman Roberto Rodes Tien Thang Pham Robert Borkowski Supannee Learkthanakhachon Gerson de Los Santos A Copenhagen based, young and dynamic team, that combines diversity in expertise and cultural backgrounds (15 nationalities) Ongoing MSc students projects (6) David Montero, visit. Asst. Prof.
    4. 4. 4 Next generation access networks services Central Office PSTN Internet Private Home with Small Repeater Mobile access Wireless Stuff Wireless Access in the City Requirements: • Versatile – handle a variety of signals • Efficient bandwidth utilization • Bidirectional • Dynamic and reconfigurable • Long-reach (~100 km)
    5. 5. 5 Hybrid fiber wireless networks CO Service integration Unified optical network platform Different modulation formats BS Different bit rates Radio over fiber (RoF) technology to increase the capacity, coverage and mobility Challenges: • Integration with existing infrastructures • Fulfill optical power budget • Increase receiver sensitivity, reach and number of users • Improve the tolerance to fiber transmission impairments • Perform signal detection and demodulation of different modulation formats and bit rates
    6. 6. 6 Where? Network scenario1 [1] Alcatel Radio-over-Fibre solution, 2007 Convergence between fixed and wireless networks the goal to bring the bandwidth of fixed network to mobile user
    7. 7. 7 To take into account: Global data traffic 1 Exabyte = 1018 bytes Drivers for traffic growth Mobility Cloud Video From CISCO analysis
    8. 8. 8 Connectivity any time, any where Source: Transfer Jet Toshiba
    9. 9. 9 Challenges Capacity Scalability and sustainability Connectivity anytime, anywhere Manageability
    10. 10. 10 100 Gbit/s wireless links Bring the capacity of baseband optical links to wireless links 1988 1992 1996 2000 2004 2008 2012 10Mbps 100Mbps 1Gbps 10Gbps 100Gbps W ireless links (standard W LAN) W ireless links (research) Optical serial interface (products) Bitrate Year Optical serial interface (research)
    11. 11. 11 How to achieve multi gigabit wireless links Higher RF carrier frequencies • GHz of bandwidth available • Higher Air attenuation Courtesy of J. Mitchell, UCL1 Frequency (GHz) 10 GSM 900MHz 1800MHz UMTS ~2GHz WLAN 2.4GHz 5.1GHz LMDS 28GHz 29GHz 31GHz HiperAccess 18GHz 42GHz MVDS 40GHz WIMAX 2.5GHz 3.5GHz 802.20 ~3.5GHz UWB 3.1-10.6 GHz Wireless HD 60GHz 6040 75 110 Future gigabit links Advanced modulation techniques • High spectral efficiency • Stringent requirements on linearity and SNR
    12. 12. 12 How to achieve multi gigabit wireless links 75-110GHz An untapped frequency band 1 100 500 Wireless HD 60GHz Saturated frequency bands (i.e GSM, UMTS, WiFi, WiMAX, WLAN, U WB, HiperAccess, LMDS, MVDS...) Frequency (GHz) 300- 500GHz  Augmented reality  HD Video Streaming  Interactive Apps 3D Skype on ipads  Mobile e-Health  Machine-to-machine  synch and go Disaster recovery Need for wireless bandwidth beyond current up to 10 GHz bands
    13. 13. 13 BTS Coverage vs Distributed Antenna Systems DAS approach •DAS attributes: •Centralization of complex equipment and simple remote antennas •Handover and load distribution/re-configurability •Power consumption BTS BTS BTS: base transceiver station
    14. 14. TU Dresden Communications Laboratory C.G. Schaeffer DTU Feb. 2009 Why Fibre Optics in Radio Systems ? numerous lowcost base stations without RF oscillators & modulators with superior RF properties higher RF carriers: - reduced cell-size - more subscribers per area - frequency-reuse - reduced RF power (EMI) low fibre attenuation for feeding the base stations remote optical generation of RF carriers broadband data signals
    15. 15. 15 Radio over fibre: basics Baseband Intermedeiate frequency (IF) up-conversion Frequency up- conversion GHz E/O conversion Light source Optical spectrum THz Optical spectrum fo fo fr fr RF spectrum Intensity modulaiton THz Optical fiber link Radio transmitter Radio-over-fibre
    16. 16. TU Dresden Communications Laboratory C.G. Schaeffer DTU Feb. 2009 Intensity Modulation & Direct Detection Optical Receiver Optical to RF Optical Fibre RF Output RF Input Optical Transmitter RF to Optical Optical Power Spectrum Optical Frequency RF-Components contributing to fRF fRF f0 Phase sensitive summation of all optically generated RF-Components at fRF
    17. 17. 17 Signal impairments E/O conversion E/O conversion Radio end Fiber dispersion Non-linearities Phase noise RF-powe fading Crosstalk Conversion efficiency Link gain Non-linearities Intermodualtion products
    18. 18. 18 Dispersion induced RF-power fading
    19. 19. TU Dresden Communications Laboratory C.G. Schaeffer DTU Feb. 2009 Intensity Modulation & Fibre Dispersion Fibre length L (km) Periodical RF power transfer: Fibre Dispersion: Wavelength: Velocity of Light: D ps km nm 17 1540nm c m s 3 108 222 cos f c DL 0 2 4 6 8 60 GHz 40 GHz 1 0.5 Opt. power Fundamental 3-dB frequency: DL c f dB 2 1 3
    20. 20. 20 Radio-over-fiber Base Station Dalma Novak
    21. 21. TU Dresden Communications Laboratory C.G. Schaeffer DTU Feb. 2009 Intensity Modulator for Sideband Modulation Principle: Mach-Zehnder Interferometer Bias points A B C Electrical voltage V (a.u.) Optical output power 0 0. 5 1 V0 A B C DC bias fMod-input Optical input Optical output Optical Spectrum f0 f1f-1 fRF = fMod Optical Spectrum f0 f1f-1 fRF = 2 fMod Optical Spectrum fop t f0 f1f-1 fRF = 4 fMod f-2 f2
    22. 22. 22 Optical Single Sideband + carrier Dalma Novak
    23. 23. 23 IF transport over fiber
    24. 24. 24 IF and remote LO transport over fiber
    25. 25. 25 Base-band-ovr fiber wireless transport
    26. 26. 26 A comparison of schemes
    27. 27. 27 Hybrid wireless-fibre systems To metropolitan network Central office Antenna base station: : Optical fibre Goal: Unified optical wireline and wireless signal transport systems Coherent detection and DSP All-optical envelope detection Approaches under study Recovery & protection Broadband wireless bridge Optical fiber link Optical fiber link
    28. 28. 28 All-optical envelope detection for wireless signals Modulation (envelope) DC Bias EAM Radio-frequency carrier Outputopticalpower - Vbias Half-wave rectified signal Lightwave carrier Base station Envelope detector Baseband data out •No High frequency mixers and oscillators •No frequency and bandwidth fixed operation For high carrier frequencies and large bandwidth reduced complexity is desirable Envelope detection with straightforward connectivity to fiber links is an interesting approach
    29. 29. 29 All-optical envelope detection Example: upstream channel EPON Desirable to use same technology for both wireline and wireless Key enabling techniques based on all-optical wireless-to-optical conversion
    30. 30. 30 Challenges/potential Why optical phase-modulation? 0 0.5 1 Transmission MZ phase (rad) 100% 0 2 4 6 0 0.5 1 1.5 2 Outputphase(rad) Drive signal (V) 600% (equivalent) Linearity: • Optical intensity modulators nonlinear  Mach-Zehnder – sinusoidal  EAM – exponential • Optical phase linear  If dominated by linear electro-optic effect Phase-modulation has no fundamental limit on the dynamic range. Large dynamic range enabling wide range of power levels
    31. 31. 31 Nonlinear and linear optical phase demodulation 0 0.5 1 Transmission MZ phase (rad) Photocurrent Signal – LO phase difference Large-signal Modulation •Open loop •Closed loop signal LO i~sin( signal - LO) • Sinusoidal response of the receiver Benefits of linear phase-modulation lost 90o optical hybrid LO Signal in I(t) Q(t) Y(t)= I(t)+jQ(t)=exp( (t))exp( (t)) Linear phase demodulation
    32. 32. 32 Converged fixed and wireless network Central office Metropolitan network CWlaser Analog-to-digital conversion RFcarrierrecovery Lineardemodulation Digitalcarrierrecovery Digital coherent receiver LO laserc Transmitter Phase modulator b Photonic wireless-wireline converged network a Carrier recovery and demodulation performed using DSP larger tolerances to phase noise and impairment compensation using DSP Same receiver structure for fixed and wireless signal detection OPSCODER project
    33. 33. 33 33 Radio over fiber (RoF) systems Phase-modulated (PM) RoF systems
    34. 34. 34 The basis: E1 OpticalHybrid Photodetectors Analog-to-Digital Converter 0 90 ELO E2 E3 E4 II(t) IQ(t) DigitalSignal Processing LO laser Optical Modulator ( )data t ES PC ES Transmitter Coherent detector Digital receiver Modulation index ( )s s pi j t data t V s sE P e Optical signal Electrical signal
    35. 35. 35 35 The basis: ELO ES PC LO laser E1 OpticalHybrid Photodetectors 0 90 E2 E3 E4 Optical signal Electrical signal ( ) cos( ( ))I out LO pi I t P P t data t V ( ) sin( ( ))Q out LO pi I t P P t data t V ( ( )) ( ) ( )pi j t data t V I Qe I t jI t Coherent receiver
    36. 36. 36 36 The basis: E1 OpticalHybrid Photodetectors 0 90 E2 E3 E4 Analog-to-Digital Converter DigitalReceiver ( ( )) ( ) ( )pi j t data t V I Qe I t jI tOptical frequency/ phase offset ( ( )) pi j t data t V e 1. DPLL 2. Linear demodulator 3. RF signal demodulation ( ( )) ln( )pi j data t V e
    37. 37. 37 37 Signal parameters estimation: Data clustering
    38. 38. 38 38 The basis:Analog-to-Digital Converter DigitalSignal Processing 1. DPLL 2. Linear demodulator 3. RF signal demodulation 0 ( ) 2 Re ( ) j t Basebanddata t S t e Complex baseband representation 0 ( ) 2 Re ( ) j t Basebanddata t S t e Frequency downconversion Quadrature demodulator Synchronizer
    39. 39. 39 39 The basis:Analog-to-Digital Converter DPLLand Lineardemodulator Quadrature demodulator Synchronizer ( ) ( ) ( )Baseband I QS t s t js t Symbol 1 Symbol 2 Symbol 3 Symbol 4 I Q
    40. 40. 40 40 The basis: problem statement Phase offset ( ) kj k Baseband ky s k e Noise
    41. 41. 41 41 Classical solution: Viterbi and Viterbi ( ) kj k Baseband ky s k e 2 ( ) ( ) ( ) ( ) l j M Baseband I Qs k s t js t e 2 ( ) ( ) ( ) ( ) l j M Baseband I Qs k s t js t e 4MQPSK 2 ( ) (2 ) ( ) 1 l j M M j lM Basebands k e e 2 ( ) (2 ) ( ) 1 l j M M j lM Basebands k e e No data (Non-data-aided)
    42. 42. 42 Classical solution: Viterbi and Viterbi 2 ( ) (2 ) ( ) 1 l j M M j lM Basebands k e e No data (Non-data-aided) ( ) kjMM M M k Baseband ky s k e Objective function( ) Re kjMM k k k L y eexp[ ]kjM M k k e j y It is maximized for one phasor ky ( )kF y arg( )ky (.) k 1 arg(.) M M ky • How to recover the phase of multi-amplitude signals? • How to estimate other data signal parameters such as modulation format? • How to track time-varying data transmission conditions?
    43. 43. 43 A novel point of view: data clustering Phase offset K-means clustering Cluster Centroid • Phase offset estimation and compensation • Reconfigurable phase offset estimation • Modulation format recognition • Frequency offset compensation
    44. 44. 44 The principle: Centroid Cluster 1 Centroid Cluster 2 Centroid Cluster 3 Centroid Cluster 4 1u 2u 3u 4u (a) Shortest distance 2u 3u 4u 1u ix 1ix (b) 1d 2d 3d 4d Updated Centroid Cluster 1 1u ix ,1newu 1ix (c)(d)
    45. 45. 45 RF phase recovery: Flexible configuration and simple upgrade for supporting different modulation formats Cluster Prototype Phase compensation Symbol 3 Symbol 4 Symbol 5 Symbol 6 Symbol 7 Symbol 8 Symbol 1 Symbol 2 Demodulation
    46. 46. 46 12 14 16 18 20 22 4 3 2 B2B, Viterbi & Viterbi B2B, k-means 40 Km, Viterbi & Viterbi 40 Km, k-means -log(BER) OSNR [dB] 1 dB 312.5 Mbaud 8PSK single carrier at 5 GHz Viterbi and Viterbi vs. K-means: K-means performs equally well as Viterbi and Viterbi
    47. 47. 47 Reconfigurable phase offset estimation: PSK QAM 16 QAM with phase offset Level threshold Level 1 Level 2 Level 3 Low complex QAM phase recovery
    48. 48. 48 Automatic modulation format detection: Level threshold Level 1 Level 2 Level 3 Multilevel detection Centroid k Cluster dmin(k,j+1) dmin(k,j) Centroid k+2 dmin(k+2,j) dmin(k+2,j+1) Centroid minvar( ( , ) )s d k j Condition of symmetry Signal Histogram/ K-means clustering Number of levels/ Number of clusters Multilevel?/ Symmetry? Right 16QAM/8PSK/ QPSK Signal Format Recognition K-means Re-initialization AMFD process Reconfigurable CarrierRecovery Wrong
    49. 49. 49 Automatic modulation format detection: Centroid k Cluster dmin(k,j+1) dmin(k,j) Centroid k+2 dmin(k+2,j) dmin(k+2,j+1) Centroid minvar( ( , ) )s d k j Condition of symmetry 100 1000 0,000 0,001 0,002 0,003 0,004 0,005 0,006 0,007 Symmetry Data samples OSNR 20 dB OSNR 24 dB Threshold QPSK 8PSK 0,000 0,005 0,010 0,015 0,020 Threshold Hypothesis Wrong QPSK id. Symmetry True QPSK id. Truedetection Wrongdetection 200 data-samples are required for automatic modulation format detection Modulation format = 8PSK
    50. 50. 50 Frequency offset compensation: The reconfigurable k-means clustering algorithm allows multifunctional tasks Frequency offset effect N samples First N/2 samples Second N/2 samples Cluster Centroid Blue constellation rotation by (1: )cleary N ( 1: 2 )darky N N
    51. 51. 51 Frequency offset compensation: The reconfigurable k-means clustering algorithm allows multifunctional tasks 12 13 14 15 16 17 18 19 20 21 22 4 3 2 Without frequency offset compensation With frequency offset compensation -log(BER) OSNR [dB] 2.7 dB 1200 1100 1000 900 800 700 600 4 3 OSNR 22 dB OSNR 20 dB -log(BER) Data-symbols / time-blocks Frequency offset 10 kHz a) b) 312.5 Mbaud 8PSK single carrier at 5 GHz
    52. 52. 52 52 Heterogeneous optical network: Reconfigurable digital coherent receiver for Metro Access Networs
    53. 53. 53 State of the art: converged WDM access link Tx. 1 Tx. 2 Tx. 3 Tx. 4 K. Prince et.al, PTL 2009 • 4 х 21.4 Gbit/s NRZ-DQPSK • 2 х 250 Mbit/s @ 5 GHz coherent Rof with phase modulation • 1 х 3.125 Gbit/s photonically generated IR-UWB • 1 х 256 QAM WiMAX @ 5.8 GHz (12 Mbaud, 70 Mbit/s) 200 GHz spacing between WDM channels NRZ-DQPSK PM RoF IR-UWB QAM-WiMAX Dedicated NRZ-DQPSK Receiver Dedicated PM RoF receiver Dedicated IR-UWB reciever Dedicated QAM-WiMAX receiver 4 dedicated receivers… Less atractive to network operators… - Maintenance & cost issues associated with Mixed receiver hardware Single reconfigurable digital photonic Receiver To support converged service delivery over a single infrastructure
    54. 54. 54 54 Experimental Details
    55. 55. 55 55 PM-OFDM Baseband VCSEL AWG AWG 78 km Deployed Fiber EDFA 20mW PPG DATA DATA / MZM /2 CW PPG DATA DATA / VOA 10 dB 1 2 4 IR-UWB AWG TLS M 3 Single Coherent Receiver 10 dBVOA 90°Optical Hybrid PwrMn LO 1 2 3 4 VCSEL DigitalPhotonic Receiver DSO Serial-Parallel Mapper IFFT CP DATAIN AWG VSG CW M 1 m wireless transmission Converged service delivery: • 5 Gbps directly modulated VCSEL • 20 Gbps QPSK baseband • 2 Gbps phase-modulated IR-UWB • 500 Mbps phase-modulated OFDM at 5 GHz carrier frequency Experimental setup
    56. 56. 56 56 PM-OFDM Baseband VCSEL AWG AWG 78 km Deployed Fiber EDFA 20mW PPG DATA DATA / MZM /2 CW PPG DATA DATA / VOA 10 dB 1 2 4 IR-UWB AWG TLS M 3 Single Coherent Receiver 10 dBVOA 90°Optical Hybrid PwrMn LO 1 2 3 4 VCSEL DigitalPhotonic Receiver DSO Serial-Parallel Mapper IFFT CP DATAIN AWG VSG CW M 1 m wireless transmission Linear demodulator IR-UWB Digital filtering (HPF) Matched filtering Symbol synchronization Signal demodulation PM-OFDM Symbol synchronization Delete Cyclic Prefix FFT Channel estimation Demapper Parallel-Serial Signal demodulation Optical frequency off-set compensation (DPLL) QPSK Baseband Equalization Clock recovery Binary decision & Differential decoding Carrier recovery IM VCSEL Timing recovery Digital filtering (HPF) Signal demodulation Thresholding Digital chromatic dispersion compensation Reconfigurable digital photonic receiver
    57. 57. 57 PM-OFDM Baseband VCSEL AWG AWG 78 km Deployed Fiber EDFA 20mW PPG DATA DATA / MZM /2 CW PPG DATA DATA / VOA 10 dB 1 2 4 IR-UWB AWG TLS M 3 Single Coherent Receiver 10 dBVOA 90°Optical Hybrid PwrMn LO 1 2 3 4 VCSEL DigitalPhotonic Receiver DSO Serial-Parallel Mapper IFFT CP DATAIN AWG VSG CW M 1 m wireless transmission Optical transmission over 78 km of deployed fiber
    58. 58. 58 Buy SmartDraw!- purchased copies print this document without a watermark. Visit www.smartdraw.com or call 1-800-768-3729.
    59. 59. 59 Results
    60. 60. 60 -26 -25 -24 -23 -22 -21 -20 5 4 3 2 B2B single channel 78km single channel -log(BER) Received Power [dBm] Coherent VCSEL(a) -26 -25 -24 -23 -22 -21 -20 5 4 3 2 B2B single channel 78km single channel 78km all wavelengths -log(BER) Received Power [dBm] Coherent VCSEL(a) No penalty for multichannel case
    61. 61. 61 -30 -29 -28 -27 -26 4 3 2 B2B single channel 78km single channel-log(BER) Received Power [dBm] (b) QPSK -30 -29 -28 -27 -26 4 3 2 B2B single channel B2B all wavelengths 78km single channel 78km all wavelengths -log(BER) Received Power [dBm] (b) QPSK 0.5 dB penalty for multichannel case
    62. 62. 62 -26 -24 -22 -20 -18 5 4 3 2 1 B2B single channel B2B all wavelengths 78km single channel 78km all wavelengths -log(BER) Received Power [dBm] (c) IR-UWB -26 -24 -22 -20 -18 5 4 3 2 1 B2B single channel 78km single channel -log(BER) Received Power [dBm] (c) IR-UWB No penalty for multichannel case
    63. 63. 63 -32 -31 -30 -29 -28 -27 5 4 3 2 B2B single channel 78 km single channel -log(BER) Received Power [dBm] (d) OFDM RoF -32 -31 -30 -29 -28 -27 5 4 3 2 B2B single channel B2B all wavelengths 78 km single channel 78 km all wavelengths -log(BER) Received Power [dBm] (d) OFDM RoF No penalty for multichannel case
    64. 64. 64 Receiver sensitivity: • -24 dBm for directly modulated VCSEL • -27 dBm for QPSK baseband • -23 dBm for phase-modulated IR-UWB • -27.5 dBm for phase-modulated OFDM
    65. 65. 65 PM-OFDM Baseband VCSEL AWG AWG 78 km Deployed Fiber EDFA 20mW PPG DATA DATA / MZM /2 CW PPG DATA DATA / VOA 10 dB 1 2 4 IR-UWB AWG TLS M 3 Single Coherent Receiver 10 dBVOA 90°Optical Hybrid PwrMn LO 1 2 3 4 VCSEL DigitalPhotonic Receiver DSO Serial-Parallel Mapper IFFT CP DATAIN AWG VSG CW M 1 m wireless transmission Summary: • Successful WDM signal demodulation for all four subsystems was demonstrated • 78 km of optical fiber transmission was achieved • A BER value below FEC threshold was achieved for all four subsystems
    66. 66. 100 Gbps Wireless Link in 75-110 GHz Band Using Photonic Technologies
    67. 67. 67 DTU Fotonik, Danmarks Tekniske Universitet Applications to gigabit wireless links • Sync and go • All wireless connectivity at business and home • HD video streaming (uncompressed) • Cloud computing • Video-calls http://wirelessgigabitalliance.org/ • Beyond LTE Cellular networks • Disaster recovery links • Fast deployment wireless networks • Extension of optical fiber links Optical fiber Optical fiber Optical fiber
    68. 68. 68 DTU Fotonik, Danmarks Tekniske Universitet Principle of RF generation by optical heterodyning •High capacity optical baseband generation •Incoherent beating of the lasers at the PD •Stringent requirement on laser linewidth •Scalable to high RF frequencies [1] U. Gliese et al., MTT 1998 [2] I. Insua et al., OFC 2009 [3] R. Sambaraju et al., PTL 2010 [4] D. Zibar et al., PTL 2011
    69. 69. 69 DTU Fotonik, Danmarks Tekniske Universitet Ƭ 16-QAM Optical Baseband Transmitter PolMux Emulator Heterodyne Upconversion PC X Y XX Y Y W-band LNA EDFA x2 LO 37 GHz 75–110 GHz 1551.6nm 1550.9 nm LO PD212.5 Gb/s PPG Ƭ 80GS/sADC Downconversion I/QSeparation TimingOffsetRecovery DecisionandBERTest Equalizer Receiver π/2 87.5 GHz 1550.9 1551.6 PD1 d X Y 36cm 6dB Ƭ 6dB Ƭ Experimental Setup •Optical baseband 16-QAM generation using binary signal generator •Free running ECL (100 kHz linewidth) as LO for photonic up-conversion •Double-stage down-conversion: 1. Electrically W-band to 1-26GHz; 2. Digitally from 1-26 GHz to baseband 16-QAM Optical Baseband Transmitter PC 1550.9 nm 12.5 Gb/s PPG Ƭ π/2 6dB Ƭ 6dB Ƭ 16-QAM Optical Baseband Transmitter PC 1550.9 nm 12.5 Gb/s PPG Ƭ π/2 6dB Ƭ 6dB Ƭ Ƭ PolMux Emulator X Y Ƭ PolMux Emulator X Y Heterodyne Upconversion XX Y Y 1551.6nm LO PD2 PD1 X Y Heterodyne Upconversion XX Y Y 1551.6nm LO PD2 PD1 X Y EDFA W-band LNA x2 LO 37 GHz 80GS/sADC Downconversion I/QSeparation TimingOffsetRecovery DecisionandBERTest Equalizer Receiver W-band LNA x2 LO 37 GHz 80GS/sADC Downconversion I/QSeparation TimingOffsetRecovery DecisionandBERTest Equalizer Receiver 87.5 GHz 1550.9 1551.6
    70. 70. 70 DTU Fotonik, Danmarks Tekniske Universitet Experiment Setup • For information: – Both signal laser and LO laser has ~100 kHz linewidth, but drifting fast within the range of 300 MHz; – Signal and LO power are set to be equal; – W-band LNA has 25 dB gain, max input power = -20 dBm; – W-band Mixer is driven by LO at 74 GHz. With input RF with frequency between 75-100 GHz, the output IF lies in the frequency range 1-26 GHz;
    71. 71. 71 DTU Fotonik, Danmarks Tekniske Universitet Experiment Setup W-band Antenna 100 GHz PD W1-WR10 Adaptor W-band Antenna W-band LNA W-band Mixer LO IF
    72. 72. 72 DTU Fotonik, Danmarks Tekniske Universitet Experiment results -1 0 1 2 3 4 5 6 7 8 9 10 5 4 3 2 1 Wireless d = 50cm Wireless d = 150cm Wireless d = 200cm -log(BER) Optical power into PD (dBm) FEC 50 Gbit/s •BER curves for 50 Gbit/s single polarization 16-QAM with different wireless distances -1 0 1 2 3 4 5 6 7 8 9 5 4 3 2 1 Wireless d = 50 cm Wireless d = 75 cm Wireless d = 120 cm -log(BER) Optical power into PD (dBm) FEC 100 Gbit/s •BER curves for 100 Gbit/s PolMux 16-QAM with different wireless distances -1 0 1 2 3 4 5 6 7 8 9 10 5 4 3 2 1 Wireless d = 50cm -log(BER) Optical power into PD (dBm) FEC 50 Gbit/s -1 0 1 2 3 4 5 6 7 8 9 10 5 4 3 2 1 Wireless d = 50cm Wireless d = 150cm -log(BER) Optical power into PD (dBm) FEC 50 Gbit/s -1 0 1 2 3 4 5 6 7 8 9 5 4 3 2 1 Wireless d = 50 cm -log(BER) Optical power into PD (dBm) FEC 100 Gbit/s -1 0 1 2 3 4 5 6 7 8 9 5 4 3 2 1 Wireless d = 50 cm Wireless d = 75 cm -log(BER) Optical power into PD (dBm) FEC 100 Gbit/s
    73. 73. 73 DTU Fotonik, Danmarks Tekniske Universitet Experiment results Ƭ 16-QAM Optical Baseband Transmitter PolMux Emulator Heterodyne Upconversion PC X Y XX Y Y W-band LNA EDFA x2 LO 37 GHz 75–110 GHz 1551.6nm 1550.9 nm LO PD212.5 Gb/s PPG Ƭ 80GS/sADC Downconversion I/QSeparation TimingOffsetRecovery DecisionandBERTest Equalizer Receiver π/2 87.5 GHz 1550.9 1551.6 PD1 d X Y 36cm 6dB Ƭ 6dB Ƭ X branch Y branch X branch Y branch Constellations for 100 Gbit/s PolMux 16-QAM signal •120 cm wireless distance •8 dBm optical power into the photodiode
    74. 74. 74 Introduction to MIMO technique 1. If all Tx antennas transmits the same data: • Increase SNR • Robust against physical disaster 2. If each Tx antenna transmits different data simultaneously: • Increase link capacity • Diversity
    75. 75. 75 Training-based MIMO channel estimation 1t ,0 T XT 2 Yt 0, T T Time X Y Training period Training period DATATX TY TX TY t1 t2 DATA DATA DATA x1 2 1 1 2 2 0 0 xx yX X Y Y xy yy RT RT T RT RT T h h h h x 1 1 2 2 1 1 2 2 xx y X X Y Yxy yy RT T RT T RT T RT T h h h h Channel transfer matrix derived Pros  Simple expression Cons Required synchronization Reduced the spectral efficiency due to the overhead
    76. 76. 76 Experimental Setup of MIMO-OFDM WDM PON with DM-VCSEL
    77. 77. 77 MIMO-OFDM WDM PON with DM-VCSEL Various separationVarious distance  Successfully demodulated below the FEC limit over 7% overhead  198.5 Mb/s net data rate with 5.65 GHz  Training symbols compensate for impairments in wireless link
    78. 78. 78 Experimental Setup of 2x2 MIMO-OFDM Fiber-Wireless transmission system based on PDM technique
    79. 79. 79 2x2 MIMO-OFDM Fiber-Wireless Transmission System Based on PDM Technique -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 5 4 3 2 1 Pol-x 22.8km SMF d=1m Pol-y 22.8km SMF d=1m Pol-x 22.8km SMF d=2m Pol-y 22.8km SMF d=2m Pol-x 22.8km SMF d=3m Pol-y 22.8km SMF d=3m -log(BER) Received optical power at PD [dBm] FEC -1 1 -1 1 -1 1 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 5 4 3 2 1 Pol-x QPSK 22.8km SMF d=1m Pol-y QPSK 22.8km SMF d=1m Pol-x 16QAM 22.8km SMF d=1m Pol-y 16QAM 22.8km SMF d=1m -log(BER) Received optical power at PD [dBm] FEC -3 -1 1 3 -3 -1 1 3 -3 -1 1 3  4-QAM-OFDM  797 Mb/s  16-QAM-OFDM  1.5 Gb/s  Training symbols compensate the optical polarization rotation and crosstalk in the wireless link
    80. 80. Find out more, videos of experiments,… metroaccessgroup idtm@fotonik.dtu.dk We look for: Researchers exchange & collaboration EU Marie Curie postdoc grant applicants, August 2013 MSc & PhD studies and research stays
    81. 81. Metro-access and short range communications group
    82. 82. 82 Conclusions of MIMO RoF • Increase link capacity • Channel estimation algorithm effectively compensate for impairments in the wireless link • VCSELs are an alternative optical source for next generation access networks • PDM alternative solution to double the capacity • High potential for future in-door networks system supporting gigabit/s wireless service

    ×