Computational Intelligence - February 2014 - 58

(a)

(b)

Figure 5 Measure of one-lead ECGs with the self-developed signal acquisition system; (a) 2
poles ECG data acquisition device-ET600, (b) data processing system.

Left
Hand
Right
Hand

ET-600

NI USB-6211

Filter

Signal Acquisition

ECG

Signal Processing and
Display

Figure 6 Operation of the acquisition system using Einthoven's Lead-I data acquisition.

Before Filtering
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0 0.2 0.4 0.6 0.8

Amplitude

Dev1/ai2

1.2 1.4 1.6 1.8
Time

2.2 2.4 2.6 2.8

Figure 7 ECG signal before filtering.

Dev1/ai2 (Filtered)

Amplitude

After Filtering
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0 0.2 0.4 0.6 0.8

1.2 1.4 1.6 1.8
Time

2.2 2.4 2.6 2.8

Figure 8 ECG signal after filtering.

IV. Experimental Results
A. ECG Acquisition

A total of 19 healthy subjects took part
in this research. Table 2 illustrates the
background information of sex, age,
height and weight of nineteen subjects.
All the test subjects' ECG signals were
collected from a self-developed handheld

feedback difference common mode signal and a buffer/balanced circuit for providing a circuit with a self common
point electrode potential. Therefore, the
first bio-potential signal can be detected
by means of the first active sensor electrode and the common point electrode.
Similarly, a second bio-potential signal
having the same magnitude but a different phase as the first bio-potential signal
can be detected by the second active
sensor electrode and the common point
electrode. The associative processing unit
for receiving the signal is processed by an
analog filter/amplifier unit which operation frequency range is from 0.5 to
40 Hz for processing or comparison of
related settings.
Fig. 5 shows the measure of ECGs
with the compact signal acquisition system. Fig. 6 shows the flow chart of the
operation of the ECG acquisition system. All the acquired ECG signals form
each subject who continuously grips the
two handles of the ET-600 for an average completion time of 1 minute. Each
acquired ECG signal was sequentially
divided into 10 seconds data length
which was aligned by their R peaks
from the full recording.
Fig. 7 shows the ECG signal acquisition in a LabVIEW instrument, collected
from the ET-600, has already been prefiltered via analogous circuits. Since the
RLC components are not accurate
enough so that the cut-off frequency is
not correct enough in the analog circuits.
To further purify the signals, a digital
low-pass filter has been used next to
remove the undesired parts. Therefore, a
fourth-order Butterworth low-pass filter
was utilized for replacement in which the
cut-off frequency is 30 Hz. Fig. 8 shows
the ECG signal after filtering.
B. Analysis of ECG

instrument ET-600 in the rest condition.
Fig. 4 shows the structure of ET-600,
which comprises a first active sensor
electrode and a second active sensor
electrode with opposite polarities; the
two active sensor electrodes connected
respectively to the pulse measure device
by conductive wires and the pulse measurement device comprises a negative

IEEE ComputatIonal IntEllIgEnCE magazInE | FEbruary 2014

The RMS level, also known as the quadratic mean, is used here as a statistical
measure of the magnitude of a varying
ECG data, which was also one of the
major indexes to differentiate between
healthy people and the people with heart
disease in many investigations of heart
rate variability. The magnitude of the
time series of ECG signal is calculated by

Table of Contents for the Digital Edition of Computational Intelligence - February 2014