University at Buffalo, USA · Prosthetics · Cable-Driven Robotics · Motion Control · Human-Centered Design

Robotic Prosthesis Emulator for User-Driven Design of Multi-DOF Transradial Prostheses

Affiliation: University at Buffalo, USA

I-PEDLE is a configurable robotic transradial prosthesis emulator built to study how wrist degrees of freedom, closed-loop control strategy, and device design affect functional upper-limb movement.

ProstheticsCable-driven designHuman testingReal-time closed-loop controlPID controlTrajectory tracking

Project overview

Highly functional upper-limb prostheses can restore dexterous motion, but adding more powered wrist functions often increases device weight, control complexity, and cost.

This project addressed the need for a configurable robotic emulator that can evaluate prosthetic wrist design choices before committing to a final device design.

My contribution

Contributed to the mechanical design, fabrication, and validation of the I-PEDLE robotic prosthesis emulator.
Supported cable-driven actuation architecture with off-board motors to keep the on-limb prosthetic module lightweight.
Worked with PID-based closed-loop position control, encoder feedback, trajectory tracking, multi-step response tests, and functional daily living tasks.
Analyzed upper-limb kinematics to understand how different wrist configurations affected compensatory shoulder and elbow movement.

Objectives

Develop a lightweight multi-DOF transradial prosthesis emulator.
Enable real-time modulation of wrist configurations.
Evaluate dynamic tracking performance and response stability.
Compare fixed and actuated wrist modes during daily living tasks.
Use parameter sweeps to study how wrist DOFs influence upper-limb coordination.

Results

The emulator achieved stable trajectory tracking across multiple frequencies and amplitudes, with tracking error remaining within 5% of the reference trajectory range.
The system showed stable multi-step response without overshoot.
During daily living tasks, actuated wrist modes reduced several compensatory shoulder motions compared with fixed wrist conditions.
The parameter sweep showed that adding more wrist DOFs did not automatically reduce compensation for every task, highlighting the importance of task-specific and user-centered prosthesis design.

Tools and methods

CAD3D PrintingCable-driven actuationOff-board motorsEncodersLoad cellsLabVIEWPID controlTrajectory trackingMotion captureBiomechanicsHuman-subject testing

Technical focus

Electromechanical system designProduct architectureReal-time closed-loop controlPID controlTrajectory trackingMotor controlEncoder feedbackSensor integrationCable-driven actuationCompact packagingTest fixturesVerification testingHuman factorsDesign validation

Project figures

The figures below summarize the I-PEDLE robotic emulator design, control architecture, experimental tasks, bench-top characterization, and human-subject task results.

Brace and cable-routing design. Upper-arm and forearm braces, fixed shoulder support, cable guides, bypass socket, and prosthesis module used to route off-board actuation cables while preserving consistent cable length during arm movement.

Mechanical design of I-PEDLE prosthesis emulator

Mechanical design of the prosthesis module. CAD renderings showing the compact 3-DOF emulator and its actuation mechanisms for pronation/supination, dart-throwing motion, and power grasp.

Full system assembly. I-PEDLE setup with off-board motors, load cells, control box, shoulder support, control interface, emergency switch, arm braces, and the lightweight on-limb prosthesis module.

Control architecture for cable-driven antagonistic motors

Control architecture. PID-based position control for antagonistic cable-driven motors, where the wrist-position error selects the driving motor and the opposing motor follows a model-based reference.

Trajectory tracking and multi-step response results

Bench-top characterization. Desired and actual encoder trajectories for pronation/supination and dart-throwing motion during sinusoidal tracking and multi-step response testing.

Daily-living task design. Table-Pour, Shelf-Place, and Shelf-Pour tasks used to evaluate how different wrist configurations affect functional manipulation and upper-limb compensation.

Table-Pour and Shelf-Place human testing results

Fixed vs. all-DOF wrist results. Range-of-motion and peak-angle comparisons during Table-Pour and Shelf-Place tasks, showing reductions in several compensatory shoulder and elbow movements with the actuated wrist.

Shelf-Pour parameter sweep results across wrist configurations

Wrist-configuration parameter sweep. Shelf-Pour results comparing fixed, DTM-only, PS-only, and all-DOF wrist modes, highlighting that prosthetic wrist benefits are task-specific and not simply proportional to the number of active DOFs.

These figures are intended for your personal academic portfolio. Use only figures you have permission to share publicly.