Main Content

OFDM with MIMO Simulation

This example shows how to use an OFDM modulator and demodulator in a simple, 2x2 MIMO error rate simulation. The OFDM parameters are based on the 802.11n standard.

Create a QPSK modulator and demodulator pair.

qpskMod = comm.QPSKModulator;
qpskDemod = comm.QPSKDemodulator;

Create an OFDM modulator and demodulator pair with user-specified pilot indices, an inserted DC null, two transmit antennas, and two receive antennas. Specify pilot indices that vary across antennas.

ofdmMod = comm.OFDMModulator('FFTLength',128, ...
    'PilotInputPort',true, ...
    'PilotCarrierIndices', ...
    cat(3,[12; 40; 54; 76; 90; 118],[13; 39; 55; 75; 91; 117]), ...
    'InsertDCNull',true, ...
ofdmDemod = comm.OFDMDemodulator(ofdmMod);
ofdmDemod.NumReceiveAntennas = 2;

Show the resource mapping of pilot subcarriers for each transmit antenna. The gray lines in the figure denote the insertion of null subcarriers to minimize pilot signal interference.


Figure OFDM Subcarrier Mapping for Tx Antenna 1 contains an axes object. The axes object with title OFDM Subcarrier Mapping for Tx Antenna 1 contains an object of type image.

Figure OFDM Subcarrier Mapping for Tx Antenna 2 contains an axes object. The axes object with title OFDM Subcarrier Mapping for Tx Antenna 2 contains an object of type image.

Determine the dimensions of the OFDM modulator by using the info method.

ofdmModDim = info(ofdmMod);

numData = ofdmModDim.DataInputSize(1);   % Number of data subcarriers
numSym = ofdmModDim.DataInputSize(2);    % Number of OFDM symbols
numTxAnt = ofdmModDim.DataInputSize(3);  % Number of transmit antennas

Generate data symbols to fill 100 OFDM frames.

nframes = 100;
data = randi([0 3],nframes*numData,numSym,numTxAnt);

Apply QPSK modulation to the random symbols and reshape the resulting column vector to match the OFDM modulator requirements.

modData = qpskMod(data(:));
modData = reshape(modData,nframes*numData,numSym,numTxAnt);

Create an error rate counter.

errorRate = comm.ErrorRate;

Simulate the OFDM system over 100 frames assuming a flat, 2x2, Rayleigh fading channel. Remove the effects of multipath fading using a simple, least squares solution, and demodulate the OFDM waveform and QPSK data. Generate error statistics by comparing the original data with the demodulated data.

for k = 1:nframes

    % Find row indices for kth OFDM frame
    indData = (k-1)*ofdmModDim.DataInputSize(1)+1:k*numData;

    % Generate random OFDM pilot symbols
    pilotData = complex(rand(ofdmModDim.PilotInputSize), ...

    % Modulate QPSK symbols using OFDM
    dataOFDM = ofdmMod(modData(indData,:,:),pilotData);

    % Create flat, i.i.d., Rayleigh fading channel 2-by-2 channel
    chGain = complex(randn(2,2),randn(2,2))/sqrt(2); 

    % Pass OFDM signal through Rayleigh and AWGN channels
    receivedSignal = awgn(dataOFDM*chGain,30);

    % Apply least squares solution to remove effects of fading channel
    rxSigMF = chGain.' \ receivedSignal.';

    % Demodulate OFDM data
    receivedOFDMData = ofdmDemod(rxSigMF.');

    % Demodulate QPSK data
    receivedData = qpskDemod(receivedOFDMData(:));

    % Compute error statistics
    dataTmp = data(indData,:,:);
    errors = errorRate(dataTmp(:),receivedData);

Display the error statistics.

fprintf('\nSymbol error rate = %d from %d errors in %d symbols\n',errors)
Symbol error rate = 9.471154e-02 from 1970 errors in 20800 symbols