Non-contact Biopotential Sensing.

摘要

Ubiquitous physiological monitoring will be a key driving force in the upcoming wireless health revolution. Cardiac and brain signals in the form of ECG and EEG are two critical health indicators that directly benefit from long-term monitoring. Despite advancements in wireless technology and electronics miniaturization, however, the use of wireless home ECG/EEG monitoring is still limited by the inconvenience and discomfort of wet, contact electrodes. This research focuses on the development of non-contact electrodes, which do not require direct electrical skin contact as a patient-friendly alternative and begins with a review of the field. Early attempts at building non-contact sensors using off-the-shelf commercial components demonstrated the feasibility of building low-cost, wireless, wearable ECG and EEG monitoring systems. As part of this early work, it was discovered that the interface noise from the insulating medium between body and sensor was often dominant, contributing significant new knowledge in this field. Further research revealed that discrete amplifiers contained many limitations, especially regarding frequency response and noise that were difficult to surmount. Previous implementations known in the literature required extensive manual tuning and calibration in order to boost the input impedance of discrete amplifiers, an imperfect and tedious process. To overcome the challenges with using discrete components, a fully custom analog sensor front-end was developed, achieving input impedances and frequency responses far exceeding than what was previously possible, all completely without the need for manual adjustment. Validation of this sensor in ECG applications show that it easily meets medical grade frequency response specifications and attains closer signal correlation to adhesive wet electrodes. Neural applications of this sensor were also explored and validated within an EEG (stead state visual evoked potential) brain-computer interface and benchmarked against dry and wet sensors. Successful real-time control of a computer, to a degree never before demonstrated with non-contact sensors, was achieved with the electrodes placed on top of hair, completely without gels or skin preparation. Additional sensor applications including EOG eye tracking and low-power integrated, focal-plane video compression are also discussed.