An open-source ESP32-S3 smartwatch that uses electrical impedance tomography to recognise hand gestures through electrodes placed on the back of the watch.
Researchers from Northwestern University have developed EITWatch, an open-source smartwatch-based gesture-sensing system built around the Seeed Studio XIAO ESP32-S3 board. Eight electrodes placed on the back of the watch detect impedance changes caused by muscle and tendon movements beneath the skin. Unlike wrist-based EIT systems that use electrodes around the wrist, EITWatch uses a flat electrode array within a standard 40-mm watch case.
The electrode configuration includes eight 2-mm-thick gold-plated stainless-steel electrodes arranged in a 31-mm-diameter ring. An analogue front end generates a 50-kHz sinusoidal excitation signal and records impedance measurements. Injection and sensing electrode pairs are connected through analogue switches. Since the flat electrode array cannot encircle the wrist, EITWatch uses multi-depth scanning to obtain 35 impedance measurements per frame at 48 Hz.
For data analysis, the ESP32-S3 runs an ESP-IDF-based programme to transmit EIT readings to a personal computer over a wireless connection. The system achieved 92.5% accuracy for six macro-gestures and 91.5% for five micro-gestures. Accuracy decreased when testing was repeated after 48 hours and dropped further when different people performed the tests.
The system can also perform gesture recognition without relying on an external device. One ESP32-S3 core handles data acquisition, while the other performs signal processing and Extra Trees classification. At 4.3 V, the system draws 35 mA, providing approximately 8.6 hours of continuous operation from a 300-mAh lithium battery.
The project is fully open source. Its Altium PCB design files, Gerber files and bill of materials are licensed under the CERN-OHL-P-2.0 licence, while its C-based firmware for the ESP32-S3 is available under the Apache-2.0 licence. The project files are hosted on GitHub.
