5GNR OFDM信号 matlab 仿真

### 5G NR OFDM Signal Simulation in MATLAB

In order to simulate a 5G New Radio (NR) Orthogonal Frequency-Division Multiplexing (OFDM) signal using MATLAB, one can follow an approach that includes setting up system parameters, generating data symbols, applying modulation, inserting cyclic prefixes, performing inverse fast Fourier transform (IFFT), adding noise, and finally analyzing the received signal. This process highlights how MATLAB serves as a powerful tool for developing and testing communication systems[^1].

Below is a simplified example demonstrating these steps:

#### System Parameters Setup
Firstly, define essential parameters such as carrier frequency, sampling rate, FFT size, number of subcarriers allocated for pilot signals, etc.

% Define basic simulation parameters
carrierFreq = 28e9;           % Carrier frequency in Hz
sampleRate = 1.92e6;          % Sampling rate in samples per second
fftSize = 2048;               % Size of IFFT/FFT operation
numPilots = round(fftSize*0.1); % Number of pilots relative to total carriers
snrRange_dB = -5:2.5:30;      % Range of SNRs to test against
monteCarloRuns = 1000;        % Number of independent runs for averaging results

#### Data Generation & Modulation
Generate random bits followed by QPSK modulation similar to what was described earlier but tailored specifically towards OFDM requirements.

dataBits = randi([0 1], fftSize-numPilots, monteCarloRuns);
modulatedData = pskmod(dataBits, 4, pi/4);    % Apply π/4 shifted QPSK

#### Pilot Insertion
Insert known reference points into specific locations within each symbol period to assist with channel estimation at receiver side.

pilotIndices = linspace(1, fftSize, numPilots+2);
pilotValues = exp(1j*pi*(rand(numPilots)-0.5));
for runIdx = 1:monteCarloRuns
    modulatedData(pilotIndices(2:end-1),runIdx) = pilotValues;
end

#### Cyclic Prefix Addition
Add guard intervals before transmitting actual payload information through wireless channels.

cpLength = floor(fftSize * 0.1);     % Length proportional to FFT length
transmittedSignal = [];
for colIndex = 1:size(modulatedData,2)
    currentSymbol = ifftshift(ifft(modulatedData(:,colIndex)));
    cpAddedSymbol = [currentSymbol(end-cpLength+1:end); currentSymbol];
    transmittedSignal = cat(1, transmittedSignal, cpAddedSymbol);
end

#### Channel Modeling & Noise Injection
Simulate multipath fading effects along with additive white Gaussian noise (AWGN).

channelCoefficients = rayleighchan(sampleRate, 70);       % Create Rayleigh channel object
noisyReceivedSignals = awgn(filter(channelCoefficients, ones(size(transmittedSignal)), sampleRate)', snrRange_dB,'measured');

#### Demodulation & Performance Analysis
Perform demodulation on received noisy symbols while estimating channels based upon inserted pilots. Calculate bit error rates across various SNR levels.

demodulatedSymbols = zeros(length(snrRange_dB),monteCarloRuns);
berResults = nan(length(snrRange_dB),1);

for idxSNR = 1:length(snrRange_dB)
    recoveredSymbols = [];
    
    for symIdx = 1:(length(noisyReceivedSignals)/fftSize)
        rxSymWithCP = reshape(noisyReceivedSignals((symIdx-1)*fftSize+(1:fftSize)+idxSNR*montecarloRuns,:,:), [], 1)';
        
        % Remove CP then apply FFT shift back again prior to taking FFT.
        strippedRxSym = circshift(rxSymWithCP(-cpLength+1:end),[-floor(cpLength/2)]);
        freqDomainRxSym = fft(strippedRxSym);
        
        estimatedChannelResp = estimate_channel(freqDomainRxSym, pilotIndices, pilotValues);
        equalizedOutput = freqDomainRxSym ./ estimatedChannelResp;
        recoveredSymbols = cat(1,recoveredSymbols,equalizedOutput);
    end
    
    demodOut = pskdemod(real(recoveredSymbols'),4,pi/4);
    [~, berResults(idxSNR)] = biterr(double(demodOut>0), double(dataBits(:)>0));
end
figure(); semilogy(snrRange_dB, berResults, '-o'); grid on;
xlabel('SNR (dB)'); ylabel('Bit Error Rate'); title('BER vs SNR Curve for 5G NR OFDM Transmission');