A Robotic System For Tracking the Ulnar Nerve

Pitt Senior Capstone

Ulnar nerve damage can cause serious motor and sensory impairments, especially impacting hand function. Accurate localization of the nerve is critical for diagnosis and treatment, but traditional methods are often manual, inconsistent, and reliant on expert judgment. This project aimed to develop a robotic system capable of tracking the ulnar nerve in real-time, improving the precision, accessibility, and consistency of clinical assessments.

We designed a compliant robotic arm integrated with surface-mounted sensors to follow the anatomical path of the ulnar nerve. I was responsible for developing the system architecture, implementing tracking control algorithms, and integrating a sensor-actuator feedback loop. The system was designed to adapt to changes in arm posture and maintain consistent nerve contact during motion. Mechanical design decisions were made to ensure both patient safety and system responsiveness.

On the left, CAD of designed Arm. Middle, Fully integrated system (Arm, Rail, Patient and Sensors). Right, resulting robot arm simulation with end effector tracked.

The prototype was tested on anatomical models and successfully tracked the nerve path within an acceptable margin of error. It remained stable and responsive under dynamic conditions, demonstrating robustness and the potential for real-world use in both clinical and rehabilitative settings.

Full system CAD

This project demonstrated the feasibility of robotic-assisted nerve tracking using sensor-based control and compliant actuation. It highlighted the value of integrating mechanical design, robotics, and control theory in solving healthcare challenges. The work strengthened my skills in control systems, sensor integration, and user-centered biomedical design. Future development will focus on implementing motorized actuation for full automation, refining the tracking algorithm with machine learning, and validating the system on human subjects.

clc; clear all; close all;

% Setup Arduino connections
a = arduino('COM9', 'Uno');
b = arduino('COM12', 'Uno');

% Initialize standard transformation matrices
syms theta X_pos Y_pos

Xrot = [1 0 0; 0 cos(theta) -sin(theta); 0 sin(theta) cos(theta)];
Yrot = [cos(theta) 0 sin(theta); 0 1 0; -sin(theta) 0 cos(theta)];
Zrot = [cos(theta) -sin(theta) 0; sin(theta) cos(theta) 0; 0 0 1];
Linear = [X_pos 0 0; 0 Y_pos 0; 0 0 0];

% Initialize system with 6 theta values and constant vectors
syms theta1 theta2 theta3 theta4 theta5 theta6

x0 = 0; y0 = 0; z0 = 0;
counter = 1;
time = 200;

% Preallocate
theta1 = zeros(200,1); theta2 = zeros(200,1); theta3 = zeros(200,1);
theta4 = zeros(200,1); theta5 = zeros(200,1); theta6 = zeros(200,1);
X_plane = zeros(200,1); theta8 = zeros(200,1);
x = zeros(200,1); y = zeros(200,1); z = zeros(200,1);

% Initial calibration
T1 = readVoltage(a, 'A0') * 1.2566;
T2 = readVoltage(a, 'A1') * 1.2566;
T3 = readVoltage(a, 'A2') * 1.2566;
T4 = readVoltage(a, 'A3') * 1.2566;
T5 = readVoltage(a, 'A4') * 1.2566;
T6 = readVoltage(a, 'A5') * 1.2566;

% Link dimensions
z1 = 6; x2 = 33; z3 = 17.5; x3 = 14; z4 = 13;
x4 = 10; z5 = 10; x5 = 8; z6 = 4;

while time > 0
    % Read voltage data and convert to degrees
    theta1(counter) = -readVoltage(a, 'A0') / 3.333 * 1.2566 - T1;
    theta2(counter) = readVoltage(a, 'A1') * 1.2566 - T2;
    theta3(counter) = readVoltage(a, 'A2') * 1.2566 - T3;
    theta4(counter) = readVoltage(a, 'A3') * 1.2566 - T4;
    theta5(counter) = readVoltage(a, 'A4') * 1.2566 - T5;
    theta6(counter) = readVoltage(a, 'A5') * 1.2566 - T6;
    
    X_plane(counter) = (readVoltage(b, 'A0') - 0.25) * (32.5 / 2); % cm
    % theta8(counter) = readVoltage(b, 'A1') * 1.2566;
    
    % Create transform matrices for each joint
    Linear = [X_plane(counter) 0 0; 0 0 0; 0 0 0];
    time = time - 1;

    H1 = [subs(Zrot, theta, theta1(counter)), [0;0;0]; 0 0 0 1];
    H2 = [subs(Yrot, theta, theta2(counter)), [0;0;z1]; 0 0 0 1];
    H3 = [subs(Yrot, theta, theta3(counter)), [x2;0;0]; 0 0 0 1];
    H4 = [subs(Zrot, theta, theta4(counter)), [0;0;z3]; 0 0 0 1];
    H5 = [subs(Yrot, theta, theta5(counter)), [x4;0;z4]; 0 0 0 1];
    H6 = [subs(Yrot, theta, theta6(counter)), [0;0;z5]; 0 0 0 1];
    H7 = [eye(3), [0;0;z6]; 0 0 0 1];

    % Get joint positions
    M1 = H1*H2;
    X1 = M1(1,4); Y1 = M1(2,4); Z1 = M1(3,4);

    M2 = M1*H3;
    X2 = M2(1,4); Y2 = M2(2,4); Z2 = M2(3,4);

    M3 = M2*H4;
    X3 = M3(1,4); Y3 = M3(2,4); Z3 = M3(3,4);

    M4 = M3*H5;
    X4 = M4(1,4); Y4 = M4(2,4); Z4 = M4(3,4);

    M5 = M4*H6;
    X5 = M5(1,4); Y5 = M5(2,4); Z5 = M5(3,4);

    M6 = M5*H7;
    X6 = M6(1,4); Y6 = M6(2,4); Z6 = M6(3,4);

    FK = H1*H2*H3*H4*H5*H6*H7;
    x(counter) = FK(1,4);
    y(counter) = FK(2,4);
    z(counter) = FK(3,4);

    % Plot
    p = plot3([x0 X1], [y0 Y1], [z0 Z1], ...
              [X1 X2], [Y1 Y2], [Z1 Z2], ...
              [X2 X3], [Y2 Y3], [Z2 Z3], ...
              [X3 X4], [Y3 Y4], [Z3 Z4], ...
              [X4 X5], [Y4 Y5], [Z4 Z5], ...
              [X5 X6], [Y5 Y6], [Z5 Z6], ...
              [x0 x(counter)], [y0 y(counter)], [z0 z(counter)]);
          
    p(1).Color = [128 128 128]/255;
    p(2).Color = [0.7 0.7 0.7];
    p(3).Color = [0.91 0.41 0.17];
    p(4).Color = [0.2 0.2 0.2];
    p(5).Color = [0.91 0.41 0.17];
    p(6).Color = [0.2 0.2 0.2];
    p(7).LineWidth = 1;
    p(7).Color = 'black';

    xlim([-70 70]);
    ylim([-70 70]);
    zlim([-10 70]);
    grid on;
    drawnow();
    pause(0.05);

    counter = counter + 1;
end