1. Fuzzy Signal Detection in Multiple-Access Ultra Wide Band Communication Systems
Advanced
Toolbox: Fuzzy Logic, Communications
Deliverables: Code .m, BER Plots, Report
π― Problem & Objective: Ultra-wide band (UWB) communications transmit extremely wide bandwidth signals at low power spectral density. In multi-access Spread Spectrum (SS) scenarios, Multiple-Access Interference (MAI) exhibits heavy non-Gaussian tails that degrade classical matched filters. Design a Mamdani Fuzzy Inference System (FIS) receiver in MATLAB to detect received pulse positions and amplitudes under severe multi-user interference, non-line-of-sight (NLOS) channels, and additive noise.
βοΈ Key MATLAB Functions:
mamfisaddInputaddOutputaddMFevalfisawgnbiterr
π Expected Output & Metrics: Bit Error Rate (BER) vs. $E_b/N_0$ curves comparing conventional correlation receiver vs. Fuzzy detector (achieving $>3.5\text{ dB}$ coding gain under 8 active users), decision surface plots, and eye diagram opening.
fis = mamfis('Name', 'UWB_Fuzzy_Detector');
fis = addInput(fis, [-2 2], 'Name', 'Correlator_Output');
fis = addMF(fis, 'Correlator_Output', 'trapmf', [-2 -2 -0.6 -0.1], 'Name', 'Bit_0');
fis = addMF(fis, 'Correlator_Output', 'gaussmf', [0.3 0], 'Name', 'Uncertain_MAI');
fis = addMF(fis, 'Correlator_Output', 'trapmf', [0.1 0.6 2 2], 'Name', 'Bit_1');
fis = addOutput(fis, [-1 1], 'Name', 'Detected_Symbol');
fis = addMF(fis, 'Detected_Symbol', 'trimf', [-1 -1 0], 'Name', 'Minus_One');
fis = addMF(fis, 'Detected_Symbol', 'trimf', [0 1 1], 'Name', 'Plus_One');
rules = ["Correlator_Output==Bit_0 => Detected_Symbol=Minus_One (1)"
"Correlator_Output==Bit_1 => Detected_Symbol=Plus_One (1)"];
fis = addRule(fis, rules);
rx_sig = [-1.4; -0.8; 0.2; 1.3; 0.9];
decisions = evalfis(fis, rx_sig);
disp('Detected Symbols:'); disp(decisions);
Est. Duration: 2β3 Weeks
Request Custom Project →
2. An Embedded Simplified Fuzzy ARTMAP Implemented on a Microcontroller for Food Classification
Advanced
Toolbox: Fuzzy Logic, Deep Learning, MATLAB Coder
Deliverables: MATLAB GUI .m, C Code, Dataset
π― Problem & Objective: Develop a portable food classification system based on potentiometric multi-electrode measurements (7 different metals) to classify floral origins of honeys. Implement a Simplified Fuzzy ARTMAP (SFA) neural-fuzzy network in a custom MATLAB GUI, optimize memory usage for 8-bit/32-bit microcontrollers, and generate standalone C code.
βοΈ Key MATLAB Functions:
fuzzyevalfisconfusionmatfitcecocpcapredict
π Expected Output & Metrics: $87.5\%+$ classification accuracy in training and testing phases, optimized memory footprint ($<16\text{ KB}$ Flash/RAM on microcontroller), and GUI parameter tuner.
X_train = rand(100, 7);
Y_train = randi([1, 4], 100, 1);
alpha = 0.001; rho_base = 0.85; max_nodes = 30;
w_weights = ones(max_nodes, 14);
A_comp = [X_train, 1 - X_train];
choice_fn = sum(min(repmat(A_comp(1,:), max_nodes, 1), w_weights), 2) ./ (alpha + sum(w_weights, 2));
[~, J_winner] = max(choice_fn);
disp(['Active SFA Prototype Node: ', num2str(J_winner)]);
Est. Duration: 2β3 Weeks
Request Custom Project →
3. Dynamic Fuzzy Controller to Meet Thermal Comfort Using Neural Network Forecasted Parameters
Intermediate
Toolbox: Fuzzy Logic, Deep Learning, Simulink
Deliverables: Code .m, Simulink .slx, Report
π― Problem & Objective: HVAC systems are non-linear multivariate systems with significant thermal lag. Design a combined neuro-fuzzy model for dynamic and automatic indoor temperature regulation. An Artificial Neural Network (ANN) forecasts future indoor temperature/humidity to feed a fuzzy logic control unit that manages compressor inverter speed and air damper angles.
βοΈ Key MATLAB Functions:
sugfisaddInputaddOutputevalfisgensurftrain
π Expected Output & Metrics: Fanger's Predicted Mean Vote (PMV) maintained within $[-0.2, +0.2]$, HVAC electrical energy consumption reduced by $18.4\%$ vs. on/off thermostat control, and dynamic membership surface visualization.
hvac_fis = sugfis('Name', 'HVAC_Thermal_Controller');
hvac_fis = addInput(hvac_fis, [-5 5], 'Name', 'Temp_Forecast_Error');
hvac_fis = addMF(hvac_fis, 'Temp_Forecast_Error', 'gaussmf', [1.2 -3], 'Name', 'Too_Cold');
hvac_fis = addMF(hvac_fis, 'Temp_Forecast_Error', 'gaussmf', [0.8 0], 'Name', 'Comfort');
hvac_fis = addMF(hvac_fis, 'Temp_Forecast_Error', 'gaussmf', [1.2 3], 'Name', 'Too_Hot');
hvac_fis = addInput(hvac_fis, [20 90], 'Name', 'Relative_Humidity');
hvac_fis = addMF(hvac_fis, 'Relative_Humidity', 'trapmf', [20 20 40 60], 'Name', 'Normal');
hvac_fis = addMF(hvac_fis, 'Relative_Humidity', 'trapmf', [50 70 90 90], 'Name', 'High');
hvac_fis = addOutput(hvac_fis, [0 100], 'Name', 'Inverter_Speed_Pct');
hvac_fis = addMF(hvac_fis, 'Inverter_Speed_Pct', 'constant', 15, 'Name', 'Low_Power');
hvac_fis = addMF(hvac_fis, 'Inverter_Speed_Pct', 'constant', 95, 'Name', 'Full_Cooling');
gensurf(hvac_fis);
Est. Duration: 1β2 Weeks
Request Custom Project →
4. Fuzzy Controlled Static Var Compensator (SVC) for Transmission Line
Intermediate
Toolbox: Fuzzy Logic, Simscape Electrical, Simulink
Deliverables: Code .m, Simscape Model, Report
π― Problem & Objective: Design and simulate a Thyristor-Controlled Reactor (TCR) with Fixed Capacitor (FC) Static Var Compensator (SVC) regulated by a Mamdani fuzzy controller on a $\lambda/8$ transmission line. Compute adaptive firing angle $\alpha$ to achieve smooth reactive power compensation at the receiving load end under heavy load variations and line faults.
βοΈ Key MATLAB Functions:
mamfistrimftrapmfevalfissimgensurf
π Expected Output & Metrics: Load bus voltage stabilization within $\pm 1\%$ of 1.0 p.u., damping of power oscillations, continuous thyristor firing angle modulation ($90^\circ \le \alpha \le 180^\circ$), and transient recovery waveforms.
svc_fis = mamfis('Name', 'SVC_Voltage_Regulator');
svc_fis = addInput(svc_fis, [-0.2 0.2], 'Name', 'Voltage_Error');
svc_fis = addMF(svc_fis, 'Voltage_Error', 'trimf', [-0.2 -0.1 0], 'Name', 'UnderVolt');
svc_fis = addMF(svc_fis, 'Voltage_Error', 'trimf', [-0.05 0 0.05], 'Name', 'Nominal');
svc_fis = addMF(svc_fis, 'Voltage_Error', 'trimf', [0 0.1 0.2], 'Name', 'OverVolt');
svc_fis = addOutput(svc_fis, [90 180], 'Name', 'Firing_Angle_Alpha');
svc_fis = addMF(svc_fis, 'Firing_Angle_Alpha', 'trimf', [90 90 135], 'Name', 'Capacitive');
svc_fis = addMF(svc_fis, 'Firing_Angle_Alpha', 'trimf', [110 135 160], 'Name', 'Floating');
svc_fis = addMF(svc_fis, 'Firing_Angle_Alpha', 'trimf', [135 180 180], 'Name', 'Inductive');
disp('SVC Fuzzy Controller Initialized Successfully');
Est. Duration: 1β2 Weeks
Request Custom Project →
5. Designing an Optimal Fuzzy Logic Controller of a DC Motor Using PSO
Intermediate
Toolbox: Fuzzy Logic, Global Optimization, Simulink
Deliverables: Code .m, Simulink .slx, PSO Convergence Plot
π― Problem & Objective: Tune the membership function parameters (shapes, center points, and base widths) of a Mamdani Fuzzy Speed Controller for a Separately Excited DC Motor using Particle Swarm Optimization (PSO). Minimize the Integral of Time-weighted Absolute Error (ITAE) and eliminate speed overshoot under rated load torque impacts.
βοΈ Key MATLAB Functions:
particleswarmmamfisevalfissimoptimizemfis
π Expected Output & Metrics: DC motor step response with zero overshoot ($M_p < 0.5\%$), settling time reduced to $<0.18\text{ s}$, ITAE metric minimized by $>42\%$ compared to classical Ziegler-Nichols PID, and PSO fitness convergence curves.
function cost = fuzzy_dc_motor_cost(params)
fis = mamfis('Name', 'PSO_DC_Fuzzy');
fis = addInput(fis, [-100 100], 'Name', 'Speed_Error');
fis = addMF(fis, 'Speed_Error', 'trimf', [-100 -params(1) 0], 'Name', 'Neg');
fis = addMF(fis, 'Speed_Error', 'trimf', [-params(2) 0 params(2)], 'Name', 'Zero');
fis = addMF(fis, 'Speed_Error', 'trimf', [0 params(1) 100], 'Name', 'Pos');
simOut = sim('dc_motor_fuzzy_model', 'ReturnWorkspaceOutputs', 'on');
e = simOut.logsout.get('error').Values.Data;
t = simOut.logsout.get('error').Values.Time;
cost = trapz(t, t .* abs(e));
end
Est. Duration: 1β2 Weeks
Request Custom Project →
6. Neuro-Fuzzy Wavelet Based Adaptive MPPT Algorithm for Photovoltaic Systems
Advanced
Toolbox: Fuzzy Logic, Wavelet, Deep Learning, Simscape
Deliverables: Code .m, Simscape PV Model, Report
π― Problem & Objective: Develop an intelligent Maximum Power Point Tracking (MPPT) controller combining fuzzy logic, neural learning, and Haar Wavelet Neural Functions (HWNF) as a gradient estimator. The controller accurately tracks the Global Maximum Power Point (GMPP) of a solar PV array under rapid solar irradiance fluctuations and non-uniform Partial Shading Conditions (PSC).
βοΈ Key MATLAB Functions:
anfisanfiseditcwtwdenoiseevalfissim
π Expected Output & Metrics: Tracking efficiency $>99.2\%$, elimination of steady-state duty cycle oscillations, and ultra-fast convergence ($<25\text{ ms}$) during step changes in solar irradiance ($400\text{ W/m}^2 \to 1000\text{ W/m}^2$).
load('pv_training_data.mat');
anfis_opt = anfisOptions('InitialFIS', 5, 'EpochNumber', 40, 'DisplayANFISInformation', 0);
fis_mppt = anfis(train_data, anfis_opt);
[cfs, frequencies] = cwt(V_pv_sampled, 'bump', 1000);
dPower_dV = gradient(P_pv) ./ (gradient(V_pv) + 1e-6);
duty_cmd = evalfis(fis_mppt, [current_irrad, current_temp, V_measured]);
Est. Duration: 2β3 Weeks
Request Custom Project →
7. Evidence-Based Uncertainty Models and PSO for Multiobjective Optimization of Engineering Systems
Advanced
Toolbox: Global Optimization, Fuzzy Logic, Statistics
Deliverables: Code .m, Pareto Frontier, Report
π― Problem & Objective: Formulate an advanced multi-objective optimization framework for engineering structural designs under aleatory and epistemic uncertainties. Integrate Dempster-Shafer Theory (DST) of evidence and Fuzzy Set Theory with Multi-Objective Particle Swarm Optimization (MOPSO) to quantify system reliability, warranty failure risk, and structural weight trade-offs.
βοΈ Key MATLAB Functions:
particleswarmfminconevalfisgamultiobjparetofront
π Expected Output & Metrics: 3D Pareto optimal front of non-dominated solutions, Belief/Plausibility uncertainty envelopes, and structural failure probability $<10^{-5}$.
nvars = 4; lb = [10 5 2 50]; ub = [50 25 10 200];
fitness_fn = @(x) [compute_mass(x), compute_warranty_risk(x)];
options = optimoptions('gamultiobj', 'PopulationSize', 60, ...
'MaxGenerations', 100, 'PlotFcn', @gaplotpareto);
[x_pareto, fval] = gamultiobj(fitness_fn, nvars, [], [], [], [], lb, ub, options);
figure; plot(fval(:,1), fval(:,2), 'ro', 'LineWidth', 1.5);
xlabel('Structural Mass (kg)'); ylabel('Uncertainty Warranty Cost ($)');
grid on; title('Evidence-Based Pareto Optimal Frontier');
Est. Duration: 3β4 Weeks
Request Custom Project →
8. Genetic Algorithm Coupled with Fuzzy Logic for Agriculture Business Systems
Beginner
Toolbox: Global Optimization, Fuzzy Logic
Deliverables: Code .m, GA Convergence Plots, Report
π― Problem & Objective: Optimize agricultural crop allocation, irrigation scheduling, and fertilizer management across multiple farm zones using Genetic Algorithms (GA) coupled with a Fuzzy Soil Suitability Evaluator to maximize economic profit while conserving scarce groundwater.
βοΈ Key MATLAB Functions:
gaoptimoptionsmamfisevalfisplotfitness
π Expected Output & Metrics: Agricultural profit increase by $26.3\%$, irrigation water conservation by $31.5\%$, and GA convergence curve across 80 generations with elite reproduction.
n_zones = 5;
lb = zeros(1, n_zones);
ub = ones(1, n_zones) * 1000;
A_eq = ones(1, n_zones); beq = 2500;
profit_fun = @(water) -sum(water .* [3.2 2.8 4.1 3.5 2.2] - 0.001*water.^2);
opts = optimoptions('ga', 'MaxGenerations', 80, 'PlotFcn', @gaplotbestf);
[opt_alloc, max_neg_profit] = ga(profit_fun, n_zones, [], [], A_eq, beq, lb, ub, [], opts);
disp(['Optimized Agricultural Profit: $', num2str(-max_neg_profit)]);
Est. Duration: 4β6 Hours
Request Custom Project →
9. Power Load Balancing Using Fuzzy Logic in Distribution Networks
Intermediate
Toolbox: Fuzzy Logic, Simscape Electrical
Deliverables: Code .m, IEEE Feeder Sim, Report
π― Problem & Objective: Alleviate three-phase unbalance and overload conditions across radial distribution feeders by designing a Fuzzy Load Transfer Controller. Analyze active and reactive branch powers and determine optimal tie-switch states and phase transfers to minimize active power losses.
βοΈ Key MATLAB Functions:
mamfisaddMFevalfisgensurfdefuzz
π Expected Output & Metrics: Phase Voltage Unbalance Rate (PVUR) reduced from $4.8\%$ to $<1.2\%$, distribution line $I^2R$ power loss reduction by $14.7\%$, and 3D rule surface showing load point transfer severity.
lb_fis = mamfis('Name', 'Feeder_Load_Balancer');
lb_fis = addInput(lb_fis, [0 150], 'Name', 'Feeder_Loading_Pct');
lb_fis = addMF(lb_fis, 'Feeder_Loading_Pct', 'trapmf', [0 0 50 75], 'Name', 'Light');
lb_fis = addMF(lb_fis, 'Feeder_Loading_Pct', 'trimf', [60 80 100], 'Name', 'Normal');
lb_fis = addMF(lb_fis, 'Feeder_Loading_Pct', 'trapmf', [90 110 150 150], 'Name', 'Overloaded');
lb_fis = addOutput(lb_fis, [0 100], 'Name', 'Transfer_Priority');
lb_fis = addMF(lb_fis, 'Transfer_Priority', 'trimf', [0 0 30], 'Name', 'Hold');
lb_fis = addMF(lb_fis, 'Transfer_Priority', 'trimf', [20 50 80], 'Name', 'Moderate');
lb_fis = addMF(lb_fis, 'Transfer_Priority', 'trimf', [70 100 100], 'Name', 'Urgent_Transfer');
Est. Duration: 1β2 Weeks
Request Custom Project →
10. Simulation of Riding a Bicycle Using Simulink & Fuzzy Cadence Controller
Beginner
Toolbox: Fuzzy Logic, Simulink
Deliverables: Simulink .slx, Code .m, Report
π― Problem & Objective: Model the longitudinal dynamics of a bicycle-rider system in Simulink, considering aerodynamic drag, rolling resistance, and hill incline. Design a dual-loop Fuzzy Controller for rider target cadence (RPM) and automatic gear shifting while avoiding gear hunting and shifting singularities.
βοΈ Key MATLAB Functions:
mamfisevalfissimset_paramplot
π Expected Output & Metrics: Rider cadence maintained smoothly at $85 \pm 5\text{ RPM}$ across fluctuating grades from $-5\%$ to $+12\%$, seamless gear selection without oscillatory shifting cycles.
cad_fis = mamfis('Name', 'Bicycle_Cadence_FIS');
cad_fis = addInput(cad_fis, [40 130], 'Name', 'Cadence_RPM');
cad_fis = addMF(cad_fis, 'Cadence_RPM', 'trapmf', [40 40 65 75], 'Name', 'Too_Slow');
cad_fis = addMF(cad_fis, 'Cadence_RPM', 'trimf', [70 85 100], 'Name', 'Optimal');
cad_fis = addMF(cad_fis, 'Cadence_RPM', 'trapmf', [95 105 130 130], 'Name', 'Too_Fast');
cad_fis = addOutput(cad_fis, [-1 1], 'Name', 'Shift_Command');
cad_fis = addMF(cad_fis, 'Shift_Command', 'trimf', [-1 -1 0], 'Name', 'Shift_Down');
cad_fis = addMF(cad_fis, 'Shift_Command', 'trimf', [-0.2 0 0.2], 'Name', 'Maintain');
cad_fis = addMF(cad_fis, 'Shift_Command', 'trimf', [0 1 1], 'Name', 'Shift_Up');
Est. Duration: 4β6 Hours
Request Custom Project →
11. Smart Farm: Automated Classifying & Grading System of Tomatoes Using Fuzzy Logic
Beginner
Toolbox: Image Processing, Fuzzy Logic
Deliverables: Code .m, Sample Images, Grading GUI
π― Problem & Objective: Develop an automated quality grading system for agricultural produce in MATLAB. Use image processing (RGB to HSV color thresholding, morphological segmentation, defect blob analysis) to extract ripeness index and defect ratio, fed into a Mamdani FIS to classify tomatoes into Grade A (Export), Grade B (Domestic), or Reject.
βοΈ Key MATLAB Functions:
rgb2hsvimbinarizeregionpropsmamfisevalfis
π Expected Output & Metrics: $96.4\%$ grading accuracy against agricultural inspection standards, processing throughput of $<45\text{ ms}$ per tomato image, and segmented color/defect masks.
img = imread('tomato_sample.jpg');
hsv_img = rgb2hsv(img);
hue_channel = hsv_img(:,:,1);
redness_score = mean(hue_channel(hue_channel < 0.1 | hue_channel > 0.9)) * 100;
defect_mask = imbinarize(rgb2gray(img), 'adaptive');
defect_area_pct = (sum(defect_mask(:)) / numel(defect_mask)) * 100;
grade_fis = mamfis('Name', 'Tomato_Grader');
quality_score = evalfis(grade_fis, [redness_score, defect_area_pct]);
disp(['Assigned Tomato Quality Grade: ', num2str(quality_score)]);
Est. Duration: 4β6 Hours
Request Custom Project →
12. Self-Tuning Fuzzy PID Controller for Non-Linear Hydraulic Actuators
Beginner
Toolbox: Fuzzy Logic, Control System, Simulink
Deliverables: Code .m, Simulink .slx, Report
π― Problem & Objective: Design a self-tuning Fuzzy PID controller where proportional ($K_p$), integral ($K_i$), and derivative ($K_d$) gains are continuously adjusted online by fuzzy inference based on tracking error $e(t)$ and change in error $\Delta e(t)$ for non-linear electrohydraulic positioners.
βοΈ Key MATLAB Functions:
mamfisevalfisgensurfsimpid
π Expected Output & Metrics: Position tracking error $<0.1\text{ mm}$, zero steady-state error under variable cylinder oil pressures, and $55\%$ reduction in settling time compared to fixed-gain PID.
fpid = mamfis('Name', 'SelfTuning_PID');
fpid = addInput(fpid, [-10 10], 'Name', 'Error');
fpid = addInput(fpid, [-5 5], 'Name', 'Delta_Error');
fpid = addOutput(fpid, [0 20], 'Name', 'Delta_Kp');
fpid = addOutput(fpid, [0 5], 'Name', 'Delta_Ki');
fpid = addOutput(fpid, [0 2], 'Name', 'Delta_Kd');
gensurf(fpid, [1 2], 1); title('Adaptive Kp Gain Surface');
Est. Duration: 4β6 Hours
Request Custom Project →
13. Fuzzy C-Means (FCM) Clustering for MRI Brain Tumor Segmentation
Intermediate
Toolbox: Fuzzy Logic, Image Processing
Deliverables: Code .m, Segmented Masks, Report
π― Problem & Objective: Segment brain MRI slices into Gray Matter (GM), White Matter (WM), Cerebrospinal Fluid (CSF), and abnormal glioblastoma tumor tissue using Fuzzy C-Means (FCM) soft clustering with spatial neighborhood regularization.
βοΈ Key MATLAB Functions:
fcmimsegfcmdicemedfilt2imoverlay
π Expected Output & Metrics: Dice similarity coefficient $>0.91$ with clinical ground truth masks, sensitivity $>94\%$, and color-coded multi-class tissue segmentations.
I = imread('brain_mri_slice.png');
I_filtered = medfilt2(rgb2gray(I), [3 3]);
options = [2.0, 100, 1e-5, false];
[centers, U] = fcm(double(I_filtered(:)), 4, options);
[~, maxU] = max(U, [], 1);
seg_img = reshape(maxU, size(I_filtered));
imshow(label2rgb(seg_img, 'jet', 'k')); title('FCM Tissue & Tumor Segmentation');
Est. Duration: 1β2 Weeks
Request Custom Project →
14. Adaptive Neuro-Fuzzy Inference System (ANFIS) for Chaotic Time-Series Prediction
Advanced
Toolbox: Fuzzy Logic, Deep Learning
Deliverables: Code .m, Prediction Plots, Report
π― Problem & Objective: Model and forecast non-linear chaotic time-series (Mackey-Glass delay differential equations and financial stock volatility) using ANFIS with hybrid backpropagation and recursive least-squares estimation.
βοΈ Key MATLAB Functions:
anfisanfisOptionsevalfisplotfiscorrcoef
π Expected Output & Metrics: Root Mean Square Error (RMSE) $<0.008$ on out-of-sample testing data, Correlation coefficient $R^2 > 0.992$, and comparative validation against classic ARIMA and MLP models.
load mackeyglass.dat;
x = mackeyglass(1:1000);
train_data = [x(1:500-18), x(1+6:500-12), x(1+12:500-6), x(1+18:500), x(1+24:500+6)];
test_data = [x(501:1000-18), x(501+6:1000-12), x(501+12:1000-6), x(501+18:1000), x(501+24:1000+6)];
opt = anfisOptions('InitialFIS', 2, 'EpochNumber', 30);
fis_model = anfis(train_data, opt);
ypred = evalfis(fis_model, test_data(:, 1:4));
rmse = sqrt(mean((ypred - test_data(:, 5)).^2));
disp(['ANFIS Test RMSE: ', num2str(rmse)]);
Est. Duration: 2β3 Weeks
Request Custom Project →
15. Fuzzy Logic-Based Collision Avoidance & Path Planning for Autonomous Mobile Robots
Intermediate
Toolbox: Fuzzy Logic, Navigation Toolbox, Simulink
Deliverables: Code .m, Navigation Sim, Report
π― Problem & Objective: Develop a real-time reactive navigation algorithm for Autonomous Mobile Robots (AMRs) in unknown dynamic warehouse environments. Process distance inputs from left, front, and right LiDAR/ultrasonic sensor sectors through a Mamdani FIS to output steering angle and linear velocity without getting trapped in local minima.
βοΈ Key MATLAB Functions:
mamfisevalfisgensurfrobotics.DifferentialDriveshow
π Expected Output & Metrics: Zero collision rate across complex obstacle layouts, smooth trajectory curvature without abrupt oscillations, and average navigation time reduction by $21\%$.
robot_fis = mamfis('Name', 'AMR_Avoidance');
robot_fis = addInput(robot_fis, [0 5], 'Name', 'Front_Distance');
robot_fis = addMF(robot_fis, 'Front_Distance', 'trapmf', [0 0 0.8 1.5], 'Name', 'Near');
robot_fis = addMF(robot_fis, 'Front_Distance', 'trapmf', [1.2 2.5 5 5], 'Name', 'Far');
robot_fis = addOutput(robot_fis, [-45 45], 'Name', 'Steering_Angle_Deg');
robot_fis = addMF(robot_fis, 'Steering_Angle_Deg', 'trimf', [-45 -30 0], 'Name', 'Hard_Left');
robot_fis = addMF(robot_fis, 'Steering_Angle_Deg', 'trimf', [-15 0 15], 'Name', 'Straight');
robot_fis = addMF(robot_fis, 'Steering_Angle_Deg', 'trimf', [0 30 45], 'Name', 'Hard_Right');
Est. Duration: 1β2 Weeks
Request Custom Project →