Inspiration
Desk workers often do not notice when their neck and shoulder muscles stay tense for hours. Timer apps remind people to take breaks, but they do not know whether the muscle actually rested. We were inspired by the idea of HR Holter monitor: measure the hidden pattern during a normal day, then act on real data.
What it does
NeckQuest is a wearable EMG “Holter” for neck tension. A surface EMG sensor placed on the left upper trapezius measures muscle activity, classifies the state as relaxed or tense, and records a workday timeline to show when the muscle did not get rest gaps.
The project includes live recognition, guided calibration, local data storage, Holter playback, and an interactive browser demo that works without hardware.
How we built it
We built the device around an ESP32-S3 board with an analog EMG sensor on ADC port. The firmware samples the signal, centers it, rectifies it, and computes a rolling RMS envelope.
A Python server provides the web UI, device connection, Holter history view, and machine-learning classifier. During calibration, the user records relaxed and tense examples. The server extracts EMG features and trains a personal SVM model. The demo and presentation are static HTML pages that can run on GitHub Pages.
Challenges we ran into
The biggest challenge was getting reliable EMG data from a noisy real-world setup. Floating ADC pins, WiFi noise, electrode placement, calibration quality.
Accomplishments that we're proud of
We are proud that NeckQuest is not just a mockup. It has firmware, a local server, a live classifier, a Holter recording concept, a guided calibration flow, research-backed reasoning, screenshots, a presentation, and an interactive demo.
We are also proud of the “measure first, then train” approach: instead of guessing when a user needs a break, NeckQuest looks at the actual muscle signal.
What we learned
We learned that EMG is very sensitive to hardware details, electrode placement, signal processing, and user-specific calibration. We also learned that a useful wearable is not only about sensors - the system needs clear instructions, trustable feedback, and a workflow that prevents meaningless data collection.
What's next for NeckQuest
Next, we want to add a second EMG channel for the right upper trapezius. This would let NeckQuest compare both shoulders during a real workday, not only detect tension on one side.
That opens a more interesting coaching path: instead of only asking "is the neck tense?", NeckQuest could show left/right imbalance, mouse-arm overload, asymmetric posture habits, and moments when one shoulder stays active while the other gets rest. Research suggests that asymmetry in upper-trapezius activation may be related to neck pain, so bilateral tracking is a natural next step for making the system more useful and personal.
Log in or sign up for Devpost to join the conversation.