EMG Gesture V2

EMG Gesture V2 is an EMG training and data-collection interface for testing gesture capture with either mock data or a live OpenBCI Ganglion signal.

Run the app

From the project root:

npm install
npm run dev

Vite will print a local URL, usually:

http://127.0.0.1:5173/

Open that URL in your browser.

Modes

The app supports two signal source modes:

• Mock: generated EMG-like signal for UI and recording-flow testing
• Connect Ganglion: live OpenBCI Ganglion signal through browser BLE

How to use Mock mode

1. Run:

npm run dev

2. Open the local Vite URL
3. Leave the source toggle on Mock
4. The graph will animate with generated signal data
5. You can test threshold-triggered recording behavior without hardware

How to use Live Ganglion mode

1. Run:

npm run dev

2. Open the local Vite URL in Chrome or Edge
3. Click Connect Ganglion
4. Choose the Ganglion from the browser Bluetooth picker
5. The live signal should begin streaming into the chart

Notes:

• Web Bluetooth works best from localhost / 127.0.0.1
• Use Chrome or Edge, not Firefox
• Make sure the Ganglion is not already connected to another app

Useful commands

Start dev server:

npm run dev

Build production bundle:

npm run build

Backend

There is a backend/ folder in the repo, but the current frontend live signal path uses direct browser BLE for the Ganglion. You do not need the backend running for the current UI workflow.

TODO

• Fixed-length recording flow: use threshold crossing only to trigger the start of capture, then record for a set duration so every sample has the same time window.
• Add a visible recording progress indicator, such as a loading bar or circular timer, while a fixed-length sample is being captured.
• Save richer sample metadata with each recording, including gesture name, timestamp, threshold used, duration, and peak signal value.
• Add a short cooldown between recordings so one long contraction does not accidentally create multiple samples.
• Improve sample quality rules so captures can be labeled more accurately as good, weak, noisy, or too short.
• Add a calibration flow for rest baseline and threshold suggestion before recording starts.
• Let the user clear all samples for the current gesture and restart collection quickly.
• Export recorded samples for training, ideally as JSON or CSV.
• Persist collected samples locally so refreshes do not wipe out a session.
• Add a lightweight session summary showing how many usable samples exist per gesture.
• Make the live status panel more explicit about source, connection state, and whether capture is idle, armed, or recording.
• Tune the UI around one-channel EMG collection so the workflow feels intentional rather than generic.