Capacitive, non-contact electrode systems for acquiring ECG, EMG, and EEG through clothing and without skin contact or conductive gel — enabling truly unobtrusive long-term health monitoring and in-vehicle driver safety applications.
Wet Ag/AgCl electrodes — the clinical standard for ECG, EEG, and EMG — require conductive gel, direct skin contact, and skilled placement. Non-contact capacitive electrodes go further: they couple to the body through one or more dielectric layers (clothing, air gaps, hair), making them compatible with fully unobtrusive embedding in garments, car seats, or small button-sized wearable modules.
This project developed the hardware, analog front-end circuit design, and motion-artifact mitigation strategies for practical non-contact biopotential systems — advancing them from bench demonstrations to wearable and automotive deployment scenarios.
Key challenge — motion artifacts: The small coupling capacitance of non-contact electrodes is highly sensitive to relative motion between the electrode and skin. Critically, the same triboelectric effect that powers harvesters corrupts biopotential signals. This project studied that crossover — characterizing, modeling, and mitigating triboelectric artifacts in non-contact wearable devices through analog feedback design and signal processing.
The non-contact sensing framework was validated across three distinct biopotential modalities, each with different amplitude, frequency, and clinical relevance:
Cardiac electrical activity for heart rate, HRV, and arrhythmia detection. Detected through up to 19 fabric layers (polyester) and in-vehicle through a car seat — without skin contact.
Brain electrical activity including alpha waves (8–12 Hz) visible during drowsy/relaxed states. Key indicator for driver alertness monitoring; detected through hair without conductive gel.
Muscle electrical activity for motion characterization and rehabilitation monitoring, acquired without traditional wet gel — enabling both upper limb and eye-blink EMG detection.
In-vehicle driver monitoring system. The foundational system demonstrated fully non-contact, nonintrusive ECG, EEG, and eye-blink detection at a distance — requiring no physical contact with the driver's skin. Validated on a high-fidelity driving simulator, the system extracted HRV and alpha-wave drowsiness indicators in real time, establishing the core capacitive front-end architecture.
Motion artifact investigation. A systematic study characterized how triboelectric charge accumulation at the electrode-skin interface degrades signal quality in non-contact wearable devices. Theoretical modeling and experimental validation provided the design guidelines — ultra-high input impedance front-end, analog DRL-style feedback loop — that inform the NCMB-Button system design.
Design and evaluation of non-contact wireless system. Extends the driver monitoring architecture to a wearable wireless form factor. Detects biopotential through fabric as thick as three-layer jeans, with performance comparable to wet electrodes. Evaluates motion artifacts across different textile materials and motion types — providing design guidelines for wearable deployment.
NCMB-Button: multi-modal wearable system. A compact (39 × 32 × 17 mm, 24 g) 3D-printed wearable module with ultra-high input impedance analog front-end, dual power supply management, and 150 mAh Li-ion battery. The feedback loop drives a driven-right-leg (DRL) style signal injection into the body to suppress common-mode noise and motion artifacts. Validated through 19 polyester layers — well beyond dry-contact electrode capability.