LabVIEW Multicore Real-Time Multi-Input Muli -Output Discrete Multitone Transceiver Testbed

LabVIEW Multicore Real-Time Multi-Input LabVIEW Multicore Real-Time Multi-Input Muli-Output Discrete Multitone Muli-Output Discrete Multitone Transceiver TestbedTransceiver Testbed

Yousof Mortazavi,Aditya Chopra, and Prof. Brian L. Evans

Wireless Networking and Communications GroupThe University of Texas at Austin

IntroductionIntroduction

Discrete Multitone Discrete Multitone ModulationModulation

DMT modulation is used in wireline communication systems (e.g. DSL)

Divide frequency selective channel into many narrowband subchannels

Transmit data over each frequency flat subchannel Modulate/demodulate multicarrier signal using Fast Fourier

Transform

MIMO DMT TestbedMIMO DMT TestbedDesign Goal: Create a 2x2 DMT hardware testbed

◦ Enable rapid prototyping/testing of new designs◦ Provide user with complete control over system

parameters◦ Connect to different cables◦ Visualize channel state and communication performance

Benefits of Hardware Testbed◦ Configure system parameters and signal processing

blocks◦ Evaluate communication performance vs.

computational complexity tradeoffs◦ Support many different cables

Design Challenges◦ Real-time constraints on transmitter and receiver

system◦ Analog front-ends for signal conditioning

Modem Implementation- Modem Implementation- HardwareHardware

TX0 TX1 RX0 RX1

EmbeddedPC

PXI-8106

PXI Backplane - PXI-1045

LPFLPF

HH

LPFLPF

HH

LPFLPF

HH

LPFLPF

HH

PXI-5421 A/D

PXI-5122 D/A

TCP Link

LPF : Low Pass FilterH: Hybrid

Modem Implementation- Modem Implementation- SoftwareSoftwareReal-Time Target

◦ LabVIEW Real-Time Vis Accesses hardware Calls DLL functions

◦ C++ Dynamic Link Library (DLL) Digital discrete-time baseband processing – Generates/processes samples sent/received

to/from hardware◦ Real-time operating system

Runs on target to guarantee real-time performance

Desktop PC◦ TCP/IP link to real-time target◦ Asynchronous visualization and control using

LabVIEW

Evolution of the TestbedEvolution of the Testbed

Bit AllocationBit Allocation Fixed amount of energy available to transmit per

DMT symbol DMT allows different number of bits transmitted on

each tone Adapt bit allocation to maximize throughput or

SNR margin on each tone Hughes Hartog bit allocation algorithm [1987]

implemented

Far-End Crosstalk Far-End Crosstalk CancellationCancellationFar End Crosstalk provides significant

deterioration in bit rate

Using vectored DMT [Ginis &Cioffi, 2002] multiple receivers operate together to cancel crosstalk

Other crosstalk cancellation methods◦ Linear: zero-forcing equalizer◦ Non-linear: successive interference

cancellation

SlicerSlicer

Vectored DMTVectored DMT

QHYQHYSuccessiveInterferenceCancellation

SuccessiveInterferenceCancellation

Estimate channel

matrix H

H = Q R

Training (per-tone)Training (per-tone)

• Uses channel estimate and both received signals to effectively cancel crosstalk

y0

y1

For each tone, H, Q and R are 2x2 matrices

Q RSymbol decoding Symbol decoding

(per-tone)(per-tone)

Experimental ResultsExperimental Results• System Parameters

• 256 tones per DMT symbol• Maximum Transmitted Voltage 5.0V• Receiver noise floor ~ -60dB

• 1000ft CAT-5 cable

• Inter-twisted pairs for maximum far-end crosstalk

• Far-end crosstalk limits SNR to ~10dB

Experimental ResultsExperimental Results

SIC – Successive Interference Cancellation

Target CPU UtilizationTarget CPU Utilization

ReferencesReferencesD. Hughes-Hartog, ”Ensemble

modem structure for imperfect transmission media.” U.S. Patents Nos. 4,679,227 (July 1987), 4,731,816 (March 1988), and 4,833,706 (May 1989)

G. Ginis and J. Cioffi, “Vectored transmission for digital subscriber line systems,” IEEE J. Select. Areas Commun., vol. 20, no. 5, pp. 1085-1104, Jun. 2002

BackupBackup

Analog Front-EndAnalog Front-End

• Hybrid circuits from Texas Instruments– Line Driver / “2-wire to 4-

wire” Interface

• Custom passive analog filters from TTE– Serve as anti-aliasing

filters for TX and RX