Project Overview
The Open Muscle Sensor Bracelet (OM-12) and Open Muscle Labeler (LASK5) are open-source tools developed to explore the potential of muscle-sensing and machine learning in prosthetic technology. Designed as educational and testing devices, these tools enable users to gain hands-on experience with muscle data capture and machine learning model training, ideal for researchers, developers, and makers interested in bio-sensing and AI applications.
OM-12 Sensor Bracelet
The OM-12 Sensor Bracelet is a wearable band equipped with a series of cells housing ESP32-C3 microcontrollers, Hall effect sensors, and mechanical keyboard switches to detect forearm muscle contractions. Each bracelet has six modular cells: three with microcontrollers and three with power management and battery circuits. The bracelet captures detailed muscle pressure changes, sending real-time data for analysis. Although not intended for direct prosthetic use, it provides a robust platform to test and refine machine learning models and prosthetic control algorithms.
Key Features:
- ESP32-C3 Microcontrollers for data processing and communication
- Dual Hall Effect Sensors in each cell for high-accuracy muscle detection
- Open-source design for easy modification and experimentation
- MicroPython Compatibility for accessible coding and customization
LASK5 Labeler System
The LASK5 Labeler is a compact device designed to label muscle data captured by the OM-12, enabling precise and customized machine learning training. Built around an ESP32-S3 microcontroller, it includes Hall effect sensors, a small display, and buttons to simplify the labeling process. This system allows users to train models in a supervised learning environment by associating specific muscle signals with finger or hand movements, providing foundational data for responsive, user-specific control applications.
Key Features:
- ESP32-S3 Microcontroller with built-in battery management
- Compact, handheld design for ease of use
- Screen and button interface for real-time data labeling
- Customizable codebase to support a variety of applications
Applications and Future Development
The OM-12 and LASK5 are designed to be modular and upgradeable, serving as a starting point for projects in:
- Prosthetic Control Prototyping: Training AI models to interpret muscle signals for adaptive responses
- Human-Machine Interfaces: Experimenting with wearable tech for robotic control
- Educational Demos: Providing a hands-on introduction to machine learning and biosensing
Get Involved!
Both the OM-12 and LASK5 are fully open-source, with design files, schematics, and a Bill of Materials available on GitHub. We welcome contributions from the community and are excited to see how others use these tools in their own projects.