Prosthetic Sensors
Open for All
Humanity
OpenMuscle builds affordable, open-source pressure myography sensors that read finger movement from the forearm. One signal, three uses: prosthetic control, a human interface device, and training data for humanoid robotics. No patents. No paywalls. Just technology that belongs to everyone.
⚠ Proof of concept — very small dataset (~683 samples). Results reflect early-stage research under optimal conditions. Larger datasets and hardware revisions are expected to improve accuracy.
What is OpenMuscle?
An open-source biosensing ecosystem for wearable prosthetic control, human interface, and robotics data capture. We use pressure myography, not EMG, to detect muscle contractions and predict individual finger movements in real time.
Open Hardware
ESP32-S3 microcontrollers, custom flexible PCBs, 3D-printed enclosures. FlexGrid packs 60 sensors into a 15×4 grid at just 3mm thick. All designs released under CERN-OHL-S v2.0.
Machine Learning
Random Forest Regressors predict finger movements from pressure data. The LASK5 labeling device captures ground-truth training data. Python pipelines process UDP sensor streams in real time.
Community Driven
Collaborators worldwide: Ultimate Robotics (EMG, Ukraine), Delta Robotics (artificial muscles), PCBWay sponsor, devEco group. Featured on Hackaday and Electromaker.
How It Works
Wear
FlexGrid wraps around the forearm with 60 pressure sensors detecting muscle topology changes
Sense
Sensors capture pressure myography data and transmit via UDP over WiFi from ESP32-S3
Learn
LASK5 captures labeled finger positions. Random Forest models train on paired sensor + label data
Predict
ML model predicts individual finger movements in real time from raw pressure readings alone
Explore the Repos
View AllOpenMuscle-FlexGrid
60-sensor (15×4) forearm band with flexible PCB and ESP32-S3. Current flagship hardware.
OpenMuscle-LASK5
Label Acquiring System K5 — handheld device for capturing ground-truth finger positions for ML training.
OpenMuscle-Software
Python & MicroPython codebase for data collection, preprocessing, and Random Forest finger prediction.
An amputee couldn't take home a prosthetic because of proprietary restrictions. That's when I decided — this technology needs to be free for everyone.
Help Us Build the Future of Prosthetics
We need testers, developers, and believers. Whether you can wear a sensor, write Python, or just spread the word — there's a place for you.