FPGA调制分类与神经网络数字预失真设计

# FPGA调制分类与神经网络数字预失真设计 ## 1. FPGA调制分类 ### 1.1 前提条件 要进行调制分类，需要以下工具和硬件： - Deep Learning Toolbox™ - Deep Learning HDL Toolbox™ - Deep Learning HDL Toolbox™ Support Package for Xilinx FPGA and SoC - Communications Toolbox™ - Xilinx™ Zynq® UltraScale+™ MPSoC ZCU102 Evaluation Kit ### 1.2 预测调制类型 训练好的卷积神经网络（CNN）可以识别以下八种数字和三种模拟调制类型： - Binary phase shift keying (BPSK) - Quadrature phase shift keying (QPSK) - 8-ary phase shift keying (8-PSK) - 16-ary quadrature amplitude modulation (16-QAM) - 64-ary quadrature amplitude modulation (64-QAM) - 4-ary pulse amplitude modulation (PAM4) - Gaussian frequency shift keying (GFSK) - Continuous phase frequency shift keying (CPFSK) - Broadcast FM (B-FM) - Double sideband amplitude modulation (DSB-AM) - Single sideband amplitude modulation (SSB-AM) 以下是相关代码： ```matlab modulationTypes = categorical(["BPSK", "QPSK", "8PSK", ... "16QAM", "64QAM", "PAM4", "GFSK", "CPFSK", ... "B-FM", "DSB-AM", "SSB-AM"]); load trainedModulationClassificationNetwork trainedNet ``` ### 1.3 生成测试信号 生成具有Rician多径衰落、中心频率和采样时间漂移以及加性高斯白噪声（AWGN）的PAM4帧，具体步骤如下： 1. 使用`randi`生成随机比特。 2. 使用`pammod`进行PAM4调制。 3. 使用`rcosdesign`设计平方根升余弦脉冲成形滤波器。 4. 使用`filter`进行脉冲成形。 5. 使用`comm.RicianChannel`应用Rician多径信道。 6. 使用`comm.PhaseFrequencyOffset`应用相位和频率偏移。 7. 使用`interp1`应用采样时间漂移。 8. 使用`awgn`添加AWGN。 以下是代码实现： ```matlab rng(123456) d = randi([0 3], 1024, 1); syms = pammod(d,4); filterCoeffs = rcosdesign(0.35,4,8); tx = filter(filterCoeffs,1,upsample(syms,8)); SNR = 30; maxOffset = 5; fc = 902e6; fs = 200e3; multipathChannel = comm.RicianChannel(... 'SampleRate', fs, ... 'PathDelays', [0 1.8 3.4] / 200e3, ... 'AveragePathGains', [0 -2 -10], ... 'KFactor', 4, ... 'MaximumDopplerShift', 4); frequencyShifter = comm.PhaseFrequencyOffset(... 'SampleRate', fs); reset(multipathChannel) outMultipathChan = multipathChannel(tx); clockOffset = (rand() * 2*maxOffset) - maxOffset; C = 1 + clockOffset / 1e6; frequencyShifter.FrequencyOffset = -(C-1)*fc; outFreqShifter = frequencyShifter(outMultipathChan); t = (0:length(tx)-1)' / fs; newFs = fs * C; tp = (0:length(tx)-1)' / newFs; outTimeDrift = interp1(t, outFreqShifter, tp); rx = awgn(outTimeDrift,SNR,0); unknownFrames = helperModClassGetNNFrames(rx); [prediction1,score1] = classify(trainedNet,unknownFrames); helperModClassPlotScores(score1,modulationTypes) ``` ### 1.4 波形生成用于训练 为每种调制类型生成10,000帧，其中80%用于训练，10%用于验证，10%用于测试。每帧长度为1024个样本，采样率为200 kHz。 ```matlab numFramesPerModType = 10000; percentTrainingSamples = 80; percentValidationSamples = 10; percentTestSamples = 10; sps = 8; spf = 1024; symbolsPerFrame = spf / sps; fs = 200e3; fc = [902e6 100e6]; ``` ### 1.5 创建信道损伤 每个帧需要经过以下信道损伤： - AWGN - Rician多径衰落 - 时钟偏移，导致中心频率偏移和采样时间漂移 ```matlab channel = helperModClassTestChannel(... 'SampleRate', fs, ... 'SNR', SNR, ... 'PathDelays', [0 1.8 3.4] / fs, ... 'AveragePathGains', [0 -2 -10], ... 'KFactor', 4, ... 'MaximumDopplerShift', 4, ... 'MaximumClockOffset', 5, ... 'CenterFrequency', 902e6) chInfo = info(channel) ``` ### 1.6 波形生成与存储 创建循环为每种调制类型生成信道受损帧，并将帧及其相应标签存储在MAT文件中。 ```matlab rng(1235) tic numModulationTypes = length(modulationTypes); channelInfo = info(channel); transDelay = 50; dataDirectory = fullfile(tempdir,"ModClassDataFiles"); disp("Data file directory is " + dataDirectory); fileNameRoot = "frame"; dataFilesExist = false; if exist(dataDirectory,'dir') files = dir(fullfile(dataDirectory,sprintf("%s*",fileNameRoot))); if length(files) == numModulationTypes*numFramesPerModType dataFilesExist = true; end end if ~dataFilesExist disp("Generating data and saving in data files...") [success,msg,msgID] = mkdir(dataDirectory); if ~success error(msgID,msg) end for modType = 1:numModulationTypes elapsedTime = seconds(toc); elapsedTime.Format = 'hh:mm:ss'; fprintf('%s - Generating %s frames\n', ... elapsedTime, modulationTypes(modType)) label = modulationTypes(modType); numSymbols = (numFramesPerModType / sps); dataSrc = helperModClassGetSource(modulationTypes(modType), sps, 2*spf, fs); modulator = helperModClassGetModulator(modulationTypes(modType), sps, fs); if contains(char(modulationTypes(modType)), {'B-FM','DSB-AM','SSB-AM'}) channel.CenterFrequency = 100e6; else channel.CenterFrequency = 902e6; end for p=1:numFramesPerModType x = dataSrc(); y = modulator(x); rxSamples = channel(y); frame = helperModClassFrameGenerator(rxSamples, spf, spf, transDelay, sps); fileName = fullfile(dataDirectory,... sprintf("%s%s%03d",fileNameRoot,modulationTypes(modType),p)); save(fileName,"frame","label") end end else disp("Data files exist. Skip data generation.") end helperModClassPlotTimeDomain(dataDirectory,modulationTypes,fs) helperModClassPlotSpectrogram(dataDirectory,modulationTypes,fs,sps) ``` ### 1.7 创建数据存储 使用`signalDatastore`对象管理生成的复波形文件。 ```matlab frameDS = signalDatastore(dataDirectory,'SignalVariableNames',["frame","label"]); ``` ### 1.8 转换复信号为实数组 将复信号转换为实值4-D数组。 ```matlab frameDSTrans = transform(frameDS,@helperModClassIQAsPages); ``` ### 1.9 划分训练、验证和测试数据 将帧划分为训练、验证和测试数据。 ```matlab splitPercentages = [percentTrainingSamples,percentValidationSamples,percentTestSamples]; [trainDSTrans,validDSTrans,testDSTrans] = helperModClassSplitData(frameDSTrans,splitPercentages); ``` ### 1.10 导入数据到内存 将文件中的数据导入内存以加快训练速度。 ```matlab pctExists = parallelComputingLicenseExists(); trainFrames = transform(trainDSTrans, @helperModClassReadFrame); rxTrainFrames = readall(trainFrames,"UseParallel",pctExists); rxTrainFrames = cat(4, rxTrainFrames{:}); validFrames = transform(validDSTrans, @helperModClassReadFrame); rxValidFrames = readall(validFrames,"UseParallel",pctExists); rxValidFrames = cat(4, rxValidFrames{:}); trainLabels = transform(trainDSTrans, @helperModClassReadLabel); rxTrainLabels = readall(trainLabels,"UseParallel",pctExists); validLabels = transform(validDSTrans, @helperModClassReadLabel); rxValidLabels = readall(validLabels,"UseParallel",pctExists); testFrames = transform(testDSTrans ```