EMBRACE | Design Projects

Summary

How EMBRACE Began

EMBRACE began as a Siemens-sponsored senior design capstone project through UC Irvine’s Biomedical Engineering program. The original project concept explored adaptive prosthetic control through biosignal processing and machine learning.

As our interdisciplinary team of biomedical engineering and computer science students formed, we brought together expertise in hardware design, signal acquisition, controls, and machine learning to refine the concept into a practical solution. This collaboration ultimately shaped EMBRACE into a portable EMG-controlled prosthetic system designed for intuitive and accessible upper-limb assistance.

Technical Approach/Methodology

How It Works

1. User Muscle Contraction

The process begins when the user contracts their right arm. These voluntary muscle contractions generate electrical signals that serve as the input for the entire EMBRACE system.

2. Mindrove EMG Armband

The Mindrove EMG armband is worn on user's right forearm and detects muscle activity non-invasively through surface electrodes. It wirelessly transmits raw EMG data to the system for processing.

3. EMG Signal Acquisition

Muscle activity is collected through the Mindrove EMG armband, allowing the system to detect user intent non-invasively. The armband captures surface electromyography signals directly from the skin, a non-invasive method.

4. Signal Processing & Feature Extraction

Raw EMG data is cleaned and processed to identify patterns related to different movements. Key features are extracted from the signal to prepare it for accurate machine learning classification.

5. Machine Learning Classification

The processed features are classified into intended hand and wrist motions using a trained machine learning model. The system recognizes multiple gesture types in real time, enabling intuitive and responsive control.

6. Microcontroller System

The classified gesture commands are sent via bluetooth to the ESP32 microcontroller, which interprets the signals and coordinates the actuation of the prosthetic hand's servo motors in real time.

7. 3D-Printed Prosthetic Hand Motion

The ESP32 control system sends commands to actuate the 3D-printed prosthetic hand, translating classified gestures into precise physical movement across 6 degrees of freedom.

Outcomes

EMBRACE is designed not just as a capstone project, but as a scalable solution with a clear path to real-world deployment. Our go-to-market strategy focuses on accessibility, clinical validation, and open-source distribution to reach the patients who need it most.

🎯 Target Market

Our primary users are individuals with transradial (below-elbow) limb loss who are seeking an affordable and intuitive prosthetic alternative. Our secondary market includes prosthetic clinics and rehabilitation centers looking for advanced, cost-effective devices for their patients.

🏥 Clinical Validation

Before broad deployment, EMBRACE will undergo clinical testing through partnerships with rehabilitation centers and prosthetic clinics. Validating real-world usability and performance is a critical step toward regulatory approval and building trust with healthcare providers.

🌐 Open Source Distribution

EMBRACE's hardware designs, software, and machine learning models will be open-sourced to allow clinics, researchers, and developers worldwide to adapt and improve the system. This approach accelerates adoption and ensures the technology reaches underserved communities globally.