Added source code

Matlab/Export_external_dataset_to_csv_train_test_random.m

@@ -0,0 +1,33 @@
+
+chunks = 1;
+version = 'v5';
+numReps = 5;
+trainPercent = 0.8;
+
+labelRepIncrement = 0; %Top avoid overwrite previous files. Default= 0
+
+pdpSizes = [3,7,15,30];
+for ii=1:1:numReps
+    for jj=1:1:length(pdpSizes)
+        pdpSize = pdpSizes(jj);
+        pathFeatures = ['./Measurements/External/RangingWithCIRData3_' version '_features_' version '_pdp_' num2str(pdpSize) '.mat'];
+
+        useNormalized=0
+        if (jj==1)
+           % Split set in training and test
+            % 50% of los/nlos in each set
+            [shuffleIndexes, trainRows, testRows ]=do_export_external_dataset_to_csv_train_test_random_v5(pathFeatures, chunks, ['Rand_' num2str(ii+labelRepIncrement) '_pdp_' num2str(pdpSize) '_'], useNormalized,trainPercent);
+        else
+            do_export_external_dataset_to_csv_train_test_random_v5(pathFeatures, chunks, ['Rand_' num2str(ii+labelRepIncrement) '_pdp_' num2str(pdpSize) '_'], useNormalized, trainPercent,shuffleIndexes, trainRows, testRows);
+        end
+
+        useNormalized=1
+        do_export_external_dataset_to_csv_train_test_random_v5(pathFeatures, chunks, ['Rand_' num2str(ii+labelRepIncrement) '_pdp_' num2str(pdpSize) '_'], useNormalized,trainPercent, shuffleIndexes,trainRows, testRows);
+    end
+end
+
+
+
+
+
+

Matlab/ExtractFeatures.m

@@ -0,0 +1,19 @@
+
+version = 'v5';
+fileName = './Measurements/External/RangingWithCIRData3_';
+
+pdp_downsampled_out = 3;
+pdp_downsampling_factor = 40;
+extractFeaturesFromCir(version, fileName, pdp_downsampled_out, pdp_downsampling_factor);
+
+pdp_downsampled_out = 7;
+pdp_downsampling_factor = 20;
+extractFeaturesFromCir(version, fileName, pdp_downsampled_out, pdp_downsampling_factor);
+
+pdp_downsampled_out = 15;
+pdp_downsampling_factor = 10;
+extractFeaturesFromCir(version, fileName, pdp_downsampled_out, pdp_downsampling_factor);
+
+pdp_downsampled_out = 30;
+pdp_downsampling_factor = 5;
+extractFeaturesFromCir(version, fileName, pdp_downsampled_out, pdp_downsampling_factor);

Matlab/do_export_external_dataset_to_csv_train_test_random_v5.m

@@ -0,0 +1,165 @@
+function [shuffleRows, rowsTrain, rowsTest]=do_export_external_dataset_to_csv_train_test_random_v5(pathFeatures, chunks, prefix, useNormalized, trainPercent, shuffleRowsExplicit, rowsTrainExplicit, rowsTestExplicit)
+
+    theFeatures = load(pathFeatures);
+
+     nlos_all = theFeatures.features.nlos;
+     rss_all = theFeatures.features.rss;
+     range_all = theFeatures.features.range;
+     energy_all = theFeatures.features.energy;
+     mean_delay_all = theFeatures.features.mean_delay;
+     rms_delay_all = theFeatures.features.rms_delay;
+     label_normalized = '';
+     if (useNormalized==1)
+         pdp_resampled_all = cell2mat(theFeatures.features.pdp_downsampled_normalized_no_noise(1:end));
+         cir_first_all = cell2mat(theFeatures.features.cir_152_normalized_no_noise(1:end));
+         cir_all = cell2mat(theFeatures.features.cir_normalized_no_noise(1:end)); 
+         label_normalized ='_normalized';
+     else
+         pdp_resampled_all = cell2mat(theFeatures.features.pdp_downsampled(1:end));
+         cir_first_all = cell2mat(theFeatures.features.cir_152(1:end));
+         cir_all = cell2mat(theFeatures.features.cir(1:end));       
+     end
+
+
+     rss_all = normalize(rss_all);
+     range_all = normalize(range_all);
+     energy_all = normalize(energy_all);
+     mean_delay_all = normalize(mean_delay_all);
+     rms_delay_all = normalize(rms_delay_all); 
+
+
+     if ~exist('shuffleRowsExplicit','var')
+         %shuffleRows = randperm(size(nlos_all,1));
+         shuffleRows =  Shuffle(1:size(nlos_all,1));
+     else
+         shuffleRows= shuffleRowsExplicit;
+     end
+
+
+     nlos_all = nlos_all(shuffleRows,:);  
+     rss_all = rss_all(shuffleRows,:);  
+     range_all = range_all(shuffleRows,:);  
+     energy_all = energy_all(shuffleRows,:);  
+     mean_delay_all =mean_delay_all(shuffleRows,:);  
+     rms_delay_all =rms_delay_all(shuffleRows,:);  
+     pdp_resampled_all = pdp_resampled_all(shuffleRows,:);  
+     cir_first_all = cir_first_all(shuffleRows,:);
+     cir_all = cir_all(shuffleRows,:);  
+
+     numMeasurements = size(nlos_all,1);
+
+
+     %Split Train and Test
+
+     if ~exist('rowsTrainExplicit','var')
+         numMeasurementsTrain = floor(numMeasurements * trainPercent);
+         numMeasurementsTrainLOS = floor(0.5*numMeasurementsTrain);
+         numMeasurementsTrainNLOS = numMeasurementsTrain - numMeasurementsTrainLOS;
+
+         secuence = 1:1:size(nlos_all,1);
+
+        losIndexes = secuence(nlos_all==0);
+        nlosIndexes = secuence(nlos_all==1);
+
+        losTrainIndexes = losIndexes(1:numMeasurementsTrainLOS);
+        nlosTrainIndexes = nlosIndexes(1:numMeasurementsTrainNLOS);
+        rowsTrain = mergesorted(losTrainIndexes, nlosTrainIndexes);
+
+        losTestIndexes = losIndexes(numMeasurementsTrainLOS+1:end);
+        nlosTestIndexes = nlosIndexes(numMeasurementsTrainNLOS+1:end);
+        rowsTest = mergesorted(losTestIndexes, nlosTestIndexes);
+     else
+         rowsTrain= rowsTrainExplicit;
+         rowsTest = rowsTestExplicit;
+     end
+
+
+     train_nlos_all = nlos_all(rowsTrain,:);  
+     train_rss_all = rss_all(rowsTrain,:);  
+     train_range_all = range_all(rowsTrain,:);  
+     train_energy_all = energy_all(rowsTrain,:);  
+     train_mean_delay_all =mean_delay_all(rowsTrain,:);  
+     train_rms_delay_all =rms_delay_all(rowsTrain,:);  
+     train_pdp_resampled_all = pdp_resampled_all(rowsTrain,:);  
+     train_cir_first_all = cir_first_all(rowsTrain,:);
+     train_cir_all = cir_all(rowsTrain,:);  
+
+     test_nlos_all = nlos_all(rowsTest,:);  
+     test_rss_all = rss_all(rowsTest,:);  
+     test_range_all = range_all(rowsTest,:);  
+     test_energy_all = energy_all(rowsTest,:);  
+     test_mean_delay_all =mean_delay_all(rowsTest,:);  
+     test_rms_delay_all =rms_delay_all(rowsTest,:);  
+     test_pdp_resampled_all = pdp_resampled_all(rowsTest,:);  
+     test_cir_first_all = cir_first_all(rowsTest,:);
+     test_cir_all = cir_all(rowsTest,:);  
+
+
+     %Saving Training
+
+     numMeasurementsTrain = size(train_nlos_all,1);
+     numMeasurementsTrainPerChunk = floor(numMeasurementsTrain/chunks);
+
+    for ii=1:1:chunks 
+
+        start = (ii-1)*numMeasurementsTrainPerChunk + 1;
+        if (ii==chunks)
+            %Last chunk
+            limit=size(train_nlos_all,1);
+        else
+            limit= (ii-1)*numMeasurementsTrainPerChunk +numMeasurementsTrainPerChunk;
+        end
+         nlos = train_nlos_all(start:limit);
+         rss = train_rss_all(start:limit);
+         range = train_range_all(start:limit);
+         energy = train_energy_all(start:limit);
+         mean_delay = train_mean_delay_all(start:limit);
+         rms_delay = train_rms_delay_all(start:limit);
+         pdp_resampled = train_pdp_resampled_all(start:limit,:);
+         cir_first = train_cir_first_all(start:limit,:);
+         cir = train_cir_all(start:limit,:);
+
+         T = table(nlos, rss, range, energy, mean_delay, rms_delay, pdp_resampled, cir_first, cir);
+         writetable(T,[prefix 'External_cir_and_pdp_set_3_TRAIN_' label_normalized '_' num2str(ii) '.csv'],'Delimiter',',','QuoteStrings',true);
+    end
+
+    %Saving Test
+
+     numMeasurementsTest = size(test_nlos_all,1);
+     numMeasurementsTestPerChunk = floor(numMeasurementsTest/chunks);
+
+
+    for ii=1:1:chunks 
+
+        start = (ii-1)*numMeasurementsTestPerChunk + 1;
+        if (ii==chunks)
+            %Last chunk
+            limit=size(test_nlos_all,1);
+        else
+            limit= (ii-1)*numMeasurementsTestPerChunk +numMeasurementsTestPerChunk;
+        end
+
+         nlos = test_nlos_all(start:limit);
+         rss = test_rss_all(start:limit);
+         range = test_range_all(start:limit);
+         energy = test_energy_all(start:limit);
+         mean_delay = test_mean_delay_all(start:limit);
+         rms_delay = test_rms_delay_all(start:limit);
+         pdp_resampled = test_pdp_resampled_all(start:limit,:);
+         cir_first = test_cir_first_all(start:limit,:);
+         cir = test_cir_all(start:limit,:);
+
+         T = table(nlos, rss, range, energy, mean_delay, rms_delay, pdp_resampled, cir_first, cir);
+         writetable(T,[prefix 'External_cir_and_pdp_set_3_TEST_' label_normalized '_' num2str(ii) '.csv'],'Delimiter',',','QuoteStrings',true);
+    end
+
+    function result = normalize(vector)
+        minValue = min(vector);
+        maxValue = max(vector);
+        result = (vector - ones(length(vector),1).*minValue)./(ones(length(vector),1).*maxValue - ones(length(vector),1).*minValue);
+    end
+
+end
+
+
+