Researchers at Northwestern University have developed EITWatch, an open-source smartwatch based on the ESP32-S3 platform. This project represents a shift in wearable technology by using Electrical Impedance Tomography to recognize hand gestures. Unlike traditional wrist-worn sensors that require a band around the entire arm, EITWatch places eight electrodes on the back of the watch case. This design choice allows the system to remain compact while still detecting muscle and tendon movements.
The hardware is built around a Seeed Studio XIAO ESP32-S3 module, which handles both data acquisition and inference. The system measures impedance changes through a specialized scanning protocol. By using one fixed source electrode and rotating the negative sink through seven other positions, the watch captures data from various depths. This configuration enables the device to interpret both macro-gestures, such as making a fist, and micro-gestures, including pinches and swipes.
Performance metrics indicate strong capability for a device of this size. In initial tests, the watch achieved 92.5 percent accuracy for macro-gestures and 91.5 percent for micro-gestures. The team also implemented an on-device processing pipeline, allowing the ESP32-S3 to perform classification without external hardware. The system runs on a 300 mAh battery, providing approximately 8.6 hours of continuous operation when performing live signal processing.
The project is fully open-source. All hardware design files are available under the CERN-OHL-P-2.0 license, and the C-based firmware is released under the Apache-2.0 license. This transparency allows other engineers to reproduce the hardware and contribute to the gesture recognition models. For those interested in the technical implementation, the design includes an analog front-end with AD5930 waveform generators and ADG738 switches to manage the current injection paths.
This development marks a change in how developers can integrate complex sensor data into small-scale wearable electronics. By avoiding the need for circumferential bands, the EITWatch suggests a viable path for gesture control in standard-sized smartwatches. All documentation and source files are accessible via the project's public repository.