Digitally Filtered ECG Signal Using Low-Cost Microcontroller PDF

Preview

Summary

Digitally Filtered ECG Signal Using Low-Cost Microcontroller Asiya M. Al-Busaidi and Lazhar Khriji Dept. of Electrical and Computer Engineering Sultan Qaboos University, Muscat, Oman [email protected] ; [email protected] Abstract—This paper demonstrates the ability of designing a required I/O featu...

Description

Accelerat ing t he world's research.

Digitally Filtered ECG Signal Using Low-Cost Microcontroller Lazhar Khriji, Asiya Al-Busaidi

Cite this paper

Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers

Download a PDF Pack of t he best relat ed papers 

Evaluat ing ECG Capt uring Using Sound-Card of PC/Lapt op

AD8232 sensor corazon dat asheet Guillermo Villalba Design and Implement at ion of a T hree-Lead Elect rocardiogram Monit oring Syst em wit h Arrhyt hmia D… Christ ian Kim Abubo, Just in Umali

Digitally Filtered ECG Signal Using Low-Cost Microcontroller Asiya M. Al-Busaidi and Lazhar Khriji Dept. of Electrical and Computer Engineering Sultan Qaboos University, Muscat, Oman [email protected] ; [email protected] Abstract—This paper demonstrates the ability of designing a low-cost, flexible and modular health-care device to sense the Electrocardiography (ECG) signal of the heart. The design is made of two sensing electrodes, analog amplifiers and a low-cost microcontroller. Digital low-pass, notch and band-pass filters were implemented because they are less expensive than analog filters. Up on its low power consumption, low cost and high performance, an Atmega32 microcontroller was used to implement all required digital filters.

required I/O features and communication modules that DSP seldom have, and easy for the beginners in this field to understand and utilize [5, 6]. The filters are usually implemented in assembly for efficiency reasons.

Keywords— ECG; digital filters; microcontroller; ADC; DAC.

The rhythm of the heart is indicated by beats per minute (bpm). The normal heart rate is about 70 bpm but lower than 60 bpm during activity is abnormal. The instantaneous heart rate could reach values as high as 200 bpm during hard exercise or athletic activity; but higher than this could be due to illness, disease, or cardiac abnormalities [2]. Thus, the proper measurements of ECG should be taken to avoid contrary results.

I. INTRODUCTION The Electrocardiography (ECG) signal represents the unique heart’s beat behavior. The characteristics of the ECG signal, including the heart rate, the PR interval, the QRS duration, the QT interval, etc., are the important evidence for doctors to diagnose diseases. Any obvious variation of the ECG waveform induced by ECG processing algorithms may cause a misdiagnosis. Thus, the target of ECG filtering is to reduce the redundancy as much as possible while to maintain clinically acceptable signal quality [1,2,3]. The electrical signal of the heart is measured with aid of skin-surface transducers called Ag/Ag-Cl electrodes. It is a bipolar low-frequency weak signal. According to the standard of the ECG institution of the USA, normal ECG signal frequencies vary in the range of 0.05-100Hz, mainly concentrate in the range of 0.05~ 35Hz and their amplitude ranges from 10μV to 5mV whose typical value is 1mV. Signals in this range of amplitudes cannot be recorder and displayed without amplification. Also, to use an ECG signal for diagnosis, it needs to be filtered to remove DC signals and artifacts as well as the high-frequency interference signals from the natural environment and the body. Many electrodes as many as ten can be used to view the electrical field of the heart from different angles. The electrodes’ signals derived from the mathematical summation, difference or average are called leads/channels. For example, from three-electrode configuration, one lead output can be generated or more depending on the circuit configuration. Therefore, this paper describes a new design of an electronic circuit to measure a single-channel ECG using two electrodes. The artifacts were canceled using digital filters designed and implemented into Atmega-32 microcontroller to reduce the analog components as much as possible. Processing of analog signals usually requires some kind of digital filtering. For extremely high filter performance, Digital Signal Processors (DSP) are usually chosen, but in many cases these are too expensive to use [4]. In these cases, Microcontrollers can be used since they are inexpensive, efficient, have all the

In this work, the designed circuit was simulated in Proteus ISIS simulator and the filters were designed with aid of Matlab and then implemented into the microcontroller in C language. II. ECG SIGNAL CHARACTERISTICS

Like any electric signal, the ECG is affected by some artifacts that should be taken in mind while designing the circuit and they are as follows:    

EMG or muscular activity. Respiration and electrodes motion. Power line interference (50/60Hz). Motion artifacts.

Thus, to design the proper filter a sample of clean ECG data was taken and artifacts were added to the signal in Matlab environment. The frequency domain analysis of the original and noisy signals was studied in the following sub-section and the proper filter design is covered in the next section. A. Frequency Domain Analysis Figure 1 shows the time domain of a typical ECG signal of 100bpm which is approximately 1.66Hz. Its frequency spectrum indicates its low frequencies. Practically, based on the power supply’s standards, the measured ECG signal is corrupted with the 60Hz/50Hz power line interference which is also amplified while measurements. To represent this noise, a 50Hz power line interference noise and Gaussian noise with 1% of the signal amplitude were generated and added to the signal. The noisy signal and its frequency spectrum are shown in Fig.1. It is clear that a 50Hz noise spectrum is correlated with the signal. Since the single-channel ECG signal is the result of subtracting the Right Arm measurement from the Left Arm measurement (Lead I = LA – RA), each electrode signal is loaded with the Gaussian noise and 50Hz power-line interference as shown in Fig. 2.

Digital converter (ADC) and a display unit which could be a PC or a microcontroller with further digital filters [5,6].

2 0

(mV)

ECG Voltage (mV)

Clean ECG Signal 4

-2 0

1

2

3

4

5

Noisy ECG Signal

4 2

2

x 10-4 Clean ECG

2 1 0 0

50

3

Time (s)

Amplitude |X(f)|

3

1

100 150 200 250

3

4

5

2 1 0 0

50 100 150 200 250

Frequency (Hz)

Frequency (Hz)

LA signal (mV)

Figure 1. The typical ECG signal with 100bpm (1.66Hz) (above), with 50Hz interference (middle), and their frequency spectrums (below).

4

Noisy Signal Clean Signal

2

 

-4 Noisy ECG x 10

|X(f)|

0

|X(f)|

-2





0

Amplitude |X(f)|

ECG Voltage (mV)

Time (s)

Here the measurements are achieved by subtracting the two electrodes attached to the patient’s upper limbs and applying one feedback electrode on the lower limb (right leg) to cancel the interference. This configuration is called one-channel or one-lead configuration using 3-electrodes. This design passes through many stages which can be divided as follows:

0

The measured signals from left arm (LA) and right arm (RA) are subtracted using instrumentation amplifier with high gain and high Common-mode rejection ratio (CMRR). An inverted amplifier is feedback to the right leg (RL) as a ground reference to cancel the interference. The output of the instrumentation amplifier is shifted above zero by a bipolar-to-unipolar converter since microcontroller cannot read negative signals. The modified output is converted into digital signal in order to be read by a digital device such as microcontroller or PC for further processing, analysis and storage.

Those points are described in details in the following subsections and illustrated in Fig. 3. A. Analog Circuit In clinical diagnoses involving the ECG signal, it is of the utmost importance that the profile of the signal be as faithfully preserved as possible. The factors affecting the quality of the recorded ECG signal are the skin-electrode-amplifier interface, electrode motion artifact, electrical interference, amplifier CMRR, amplifier frequency response, semiconductor noise generated in the amplifier, and input signal level variation. In this paper, the CMRR is considered [1].

-2 0

1

2

3

4

5

RA Signal (mV)

Time (s)

Instrumentation Amplifier Bipolar-to-Unipolar LA Converter

4 Noisy Signal Clean Signal

2

RA

0 -2 -4

0

1

2

3

4

5

RL

Time (s) Figure 2. The ECG signal of the Left Arm (LA) and Right Arm (RA) with noise and without noise.

Microcontroller NF

LP

Figure 3. The proposed ECG circuit.

ADC

III. PROPOSED CIRCUIT DESIGN The typical single-channel ECG circuit consists of an Instrumentation Amplifier (IA), common-mode feedback amplifier (ground or reference electrode), Low-Pass filter (LPF), High-Pass filter (HPF), Notch filter (NF), Analog-to-

Filtered Output

DAC

BP...