Mini MATLAB Project Ideas (Practical Ideas with Complete Code)

% 1. Define Mini-CNN Architecture for Gender Classification
layers = [
    imageInputLayer([64 64 3], 'Name', 'input')
    convolution2dLayer(3, 16, 'Padding', 'same', 'Name', 'conv_1')
    batchNormalizationLayer('Name', 'bn_1')
    reluLayer('Name', 'relu_1')
    maxPooling2dLayer(2, 'Stride', 2, 'Name', 'pool_1')
    
    convolution2dLayer(3, 32, 'Padding', 'same', 'Name', 'conv_2')
    batchNormalizationLayer('Name', 'bn_2')
    reluLayer('Name', 'relu_2')
    maxPooling2dLayer(2, 'Stride', 2, 'Name', 'pool_2')
    
    fullyConnectedLayer(64, 'Name', 'fc_1')
    reluLayer('Name', 'relu_fc')
    dropoutLayer(0.4, 'Name', 'drop_1')
    fullyConnectedLayer(2, 'Name', 'fc_gender')
    softmaxLayer('Name', 'softmax')
    classificationLayer('Name', 'output')
];

% 2. Training Hyperparameters
opts = trainingOptions('adam', ...
    'InitialLearnRate', 1e-3, ...
    'MaxEpochs', 15, ...
    'MiniBatchSize', 64, ...
    'Shuffle', 'every-epoch', ...
    'Plots', 'training-progress', ...
    'Verbose', false);

% 3. Inference Demonstration
testFace = imread('sample_face.jpg');
testFaceResized = imresize(testFace, [64 64]);
[predGender, scores] = classify(net, testFaceResized);
fprintf('Predicted Gender: %s (Confidence: %.2f%%)\n', string(predGender), max(scores)*100);

% 1. Synthesize High-Dimensional CAD Feature Matrix (N samples, D=181 features)
N = 250; D = 181; k = 20; % Reduce from 181 to 20 dimensions
X_high = zscore(randn(N, D));
labels = [ones(125, 1); -ones(125, 1)]; % Malignant (+1) vs Benign (-1)

% 2. Construct Achlioptas / Gaussian Random Projection Matrix W
W = (1/sqrt(k)) * randn(D, k); % Size: [181 x 20]
X_projected = X_high * W;      % Low-dimensional feature matrix

% 3. Train Support Vector Machine with 10-Fold Cross-Validation
svmModel = fitcsvm(X_projected, labels, 'KernelFunction', 'rbf', 'Standardize', true);
cvSVM = crossval(svmModel, 'KFold', 10);
cvLoss = kfoldLoss(cvSVM);
[~, score] = kfoldPredict(cvSVM);

% 4. ROC Curve Analysis
[X_roc, Y_roc, ~, AUC] = perfcurve(labels, score(:,2), 1);
fprintf('10-Fold Cross-Validation Accuracy: %.2f%%\n', (1 - cvLoss)*100);
fprintf('Optimized RPA-SVM Area Under Curve (AUC): %.4f\n', AUC);

% 1. Read & Preprocess Document / License Plate Image
img = imread('scanned_receipt.png');
if size(img, 3) == 3, gray = rgb2gray(img); else, gray = img; end

% Adaptive Binarization and Noise Cleanup
bw = imbinarize(gray, 'adaptive', 'Sensitivity', 0.55);
bwClean = bwareaopen(~bw, 15); % Remove small pixel artifacts

% 2. Perform Optical Character Recognition (OCR)
ocrResults = ocr(gray, 'TextLayout', 'Block');

% 3. Highlight Recognized Words with Confidence Bounding Boxes
annotatedImg = insertObjectAnnotation(img, 'rectangle', ...
    ocrResults.WordBoundingBoxes, ocrResults.Words, 'FontSize', 14);

figure; imshow(annotatedImg);
title('OCR Detected Text & Bounding Regions');
fprintf('Extracted Text Content:\n%s\n', ocrResults.Text);

% 1. Deploy Sensor Nodes along Sewer Network
numNodes = 20;
nodeX = rand(1, numNodes) * 500; % 500m x 500m urban grid
nodeY = rand(1, numNodes) * 500;
sinkPos = [250, 250];            % Central Control Station

% 2. Simulate Sewer Water Level Readings & Threshold Trigger
waterLevels = 20 + 75*rand(1, numNodes); % cm
threshold = 70;                          % 70 cm Alert Limit
alertNodes = find(waterLevels > threshold);

% 3. Compute Multihop Distance & Energy Consumption
dists = pdist2([nodeX', nodeY'], sinkPos);
packetSuccess = exp(-dists / 300); % Zigbee path loss model

figure; scatter(nodeX, nodeY, 80, waterLevels, 'filled'); hold on;
plot(sinkPos(1), sinkPos(2), 'kp', 'MarkerSize', 15, 'MarkerFaceColor', 'r');
colorbar; title('Zigbee Sewerage Sensor Telemetry & Alerts');
xlabel('X (m)'); ylabel('Y (m)');

% 1. Initialize Foreground Detector for Vehicle Tracking
detector = vision.ForegroundDetector('NumGaussians', 3, 'NumTrainingFrames', 40);
blobAnalysis = vision.BlobAnalysis('BoundingBoxOutputPort', true, ...
    'AreaOutputPort', true, 'MinimumBlobArea', 250);

% 2. Process Simulated Traffic Frame
frame = imread('traffic_intersection.jpg');
fgMask = detector(frame);
[areas, bboxes] = blobAnalysis(fgMask);

% 3. Classify Vehicles (Cars vs Heavy Trucks) & Compute Green Time
numCars = sum(areas >= 250 & areas < 1200);
numTrucks = sum(areas >= 1200);
trafficDensity = numCars * 1.0 + numTrucks * 2.5;

% Dynamic Signal Timing Equation
greenTimeSeconds = min(max(15, round(trafficDensity * 3.5)), 60);
fprintf('Vehicles Detected: %d Cars, %d Trucks | Green Light Time: %d sec\n', ...
    numCars, numTrucks, greenTimeSeconds);

% State Machine Logic for Bi-Directional IR Beam Sensors
persistent count state;
if isempty(count), count = 0; state = 'IDLE'; end

% Simulated Sensor Inputs (1 = Beam Broken, 0 = Clear)
sensorA = 1; sensorB = 0; % Outer & Inner Beams

switch state
    case 'IDLE'
        if sensorA && ~sensorB, state = 'ENTERING';
        elseif ~sensorA && sensorB, state = 'EXITING'; end
    case 'ENTERING'
        if ~sensorA && sensorB
            count = count + 1;
            state = 'IDLE';
            fprintf('Person Entered. Current Occupancy: %d\n', count);
        end
    case 'EXITING'
        if sensorA && ~sensorB
            count = max(0, count - 1);
            state = 'IDLE';
            fprintf('Person Exited. Current Occupancy: %d\n', count);
        end
end

% 1. Establish Serial RF Telemetry Connection
s = serialport("COM4", 9600, "Timeout", 1);
configureTerminator(s, "CR/LF");

% 2. Telemetry Polling Loop
metalThreshold = 750; % Analog Inductive Sensor Threshold (0-1023)
disp('Robotic Telemetry Active. Monitoring Proximity...');

for k = 1:50
    if s.NumBytesAvailable > 0
        dataStr = readline(s);
        telemetry = str2double(strsplit(dataStr, ','));
        metalVal = telemetry(1); distVal = telemetry(2);
        
        if metalVal > metalThreshold
            fprintf('[ALERT] Ferromagnetic Metallic Material Detected! Val: %d\n', metalVal);
            beep; % Trigger Audio Warning
            write(s, "STOP_MOTORS", "string"); % Safety Interlock
        end
    end
    pause(0.1);
end

% 1. Sensor Calibration & Voltage Reading
rawVoltage = 2.4; % Simulated MQ-3 sensor voltage (0 - 5.0 V)
bacLevelPPM = (rawVoltage / 5.0) * 1000; % Calibration to PPM
legalLimitPPM = 350; % Legal Safety Threshold

% 2. Decision Logic & Relay Interlock
if bacLevelPPM >= legalLimitPPM
    engineRelayState = 0; % Cut Ignition Power
    statusMsg = 'Drunk Driving Detected! Engine Ignition LOCKED.';
    alarmTone = true;
else
    engineRelayState = 1; % Enable Normal Ignition
    statusMsg = 'Driver Sober. Engine Start Permitted.';
    alarmTone = false;
end

fprintf('Sensor Voltage: %.2fV | BAC: %.1f PPM | %s\n', ...
    rawVoltage, bacLevelPPM, statusMsg);

% 1. Multichannel Token FIFO Queue Data Structure
numCounters = 4;
queue = 101:125; % Incoming Token Numbers
counterStatus = repmat("IDLE", 1, numCounters);

% 2. Dispatch Next Token to Available Counter
function CallNextCustomer(counterIdx)
    global queue counterStatus;
    if ~isempty(queue)
        currentToken = queue(1);
        queue(1) = []; % Pop from FIFO queue
        counterStatus(counterIdx) = sprintf("Serving Token %d", currentToken);
        
        % Audio Announcement String
        callStr = sprintf('Token Number %d, please proceed to Counter %d', currentToken, counterIdx);
        disp(callStr);
        % Synthesize speech if available
        if exist('tts', 'file'), tts(callStr); end
    end
end

% 1. Synthesize DTMF Tone for Candidate Key '3' (697 Hz + 1477 Hz)
fs = 8000; t = 0:1/fs:0.25; % 250 ms duration
dtmfSignal = sin(2*pi*697*t) + sin(2*pi*1477*t) + 0.15*randn(size(t));

% 2. Standard DTMF Frequencies Definition
lowFreqs = [697 770 852 941];
highFreqs = [1209 1336 1477 1633];

% 3. Goertzel Algorithm for Fast Spectral Energy Extraction
lowIdx = round(lowFreqs/fs * length(dtmfSignal)) + 1;
highIdx = round(highFreqs/fs * length(dtmfSignal)) + 1;

magLow = abs(goertzel(dtmfSignal, lowIdx));
[~, detLow] = max(magLow);
magHigh = abs(goertzel(dtmfSignal, highIdx));
[~, detHigh] = max(magHigh);

keypad = ['1' '2' '3' 'A'; '4' '5' '6' 'B'; '7' '8' '9' 'C'; '*' '0' '#' 'D'];
votedCandidate = keypad(detLow, detHigh);
fprintf('Decoded Candidate Vote: %s (Tone Freqs: %d Hz, %d Hz)\n', ...
    votedCandidate, lowFreqs(detLow), highFreqs(detHigh));

% 1. Authorized User Database (Card UID -> Encrypted SHA-256 PIN Hash)
userDB = containers.Map();
userDB('CARD_84920') = '8c6976e5b5410415bde908bd4dee15dfb167a9c873fc4bb8a81f6f2ab448a918'; % PIN: 'admin'

% 2. Authentication Function
cardUID = 'CARD_84920';
inputPIN = 'admin';

if isKey(userDB, cardUID)
    % Generate SHA-256 Hash of Entered PIN
    md = java.security.MessageDigest.getInstance('SHA-256');
    hashBytes = md.digest(double(inputPIN));
    hashedPIN = sprintf('%02x', typecast(int8(hashBytes), 'uint8'));
    
    if strcmp(hashedPIN, userDB(cardUID))
        disp('[SUCCESS] Access Granted! Door Solenoid Unlocked for 5s.');
        accessGranted = true;
    else
        disp('[DENIED] Invalid PIN. Security Warning Logged.');
        accessGranted = false;
    end
else
    disp('[DENIED] Unregistered Smart Card UID.');
end