Long-Term Polygraphic Monitoring through MEMS and Charge Transfer for Low-Power Wearable Applications

Abstract

In this work, we propose a wireless wearable system for the acquisition of multiple biopotentials through charge transfer electrostatic sensors realized in MEMS technology. The system is designed for low power consumption and low invasiveness, and thus candidates for long-time monitoring in free-living conditions, with data recording on an SD or wireless transmission to an external elaborator. Thanks to the wide horizon of applications, research is very active in this field, and in the last few years, some devices have been introduced on the market. The main problem with those devices is that their operation is time-limited, so they do not match the growing demand for long monitoring, which is a must-have feature in diagnosing specific diseases. Furthermore, their versatility is hampered by the fact that they have been designed to record just one type of signal. Using ST-Qvar sensors, we acquired an electrocardiogram trace and single-channel scalp electroencephalogram from the frontal lobes, together with an electrooculogram. Excellent results from all three types of acquisition tests were obtained. The power consumption is very low, demonstrating that, thanks to the MEMS technology, a continuous acquisition is feasible for several days.

Keywords: MEMS technology; electrostatic sensors; long time domestic monitoring; low power consumption; multiple biopotentials acquisition; wearable sensors.

Figure 1

(a) The Nucleo L476RG + IKS01A1, (b) the ST-Qvar board, and (c) the Bluetooth module.

Figure 2

Pre-amplifier circuit.

Figure 3

(a) The detailed high−pass filter; (b) gain vs. frequency.

Figure 4

Simulations of the differential stage performed on LT Spice: (a) Differential gain; (b) common−mode gain.

Figure 5

Noise analysis of the proposed pre-amp stage with Zep = ∞.

Figure 6

Electrode model according to IEC 60601–2–47 standard.

Figure 7

Noise analysis of the proposed pre-amp stage with Zep = 0.

Figure 8

Monte Carlo simulations on pre−amplifier stage with 1% passive tolerance.

Figure 9

Matlab application interface for data acquisition.

Figure 10

Block diagram of the proposed firmware. The overall system clock is 3 MHz and the interrupt clock is 240 Hz.

Figure 11

(a) Electrodes positioning; (b) typical ECG raw trace.

Figure 12

(a) Filtered ST−Qvar ECG: single oscillation; (b) Typical ECG, single oscillation.

Figure 13

ECG trace acquired simultaneously from Qvar and MicroMed systems, from different derivations. Dashed lines show the R−peaks correspondence between the two tests.

Figure 14

Electrodes positioning for EEG–EOG test for MicroMed (left, orange electrodes) and ST-Qvar (right, blue electrodes).

Figure 15

MicroMed vs. ST−Qvar raw traces from frontal single-channel EEG acquisition.

Figure 16

Cross-correlation between ST−Qvar and MicroMed acquisitions. MicroMed signal autocorrelation.

Figure 17

EOG trace recorded from ST−Qvar and the gold standard during clockwise eye movements. A single window of the eye movements is highlighted with a rectangle.

Figure 18

EOG trace recorded from ST−Qvar and the gold standard during eye blinking test. Each peak represents an eye blink.