See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/342872166
ASIC Implementation of an Adaptive Noise Canceller for ECG Signal Processing
Applications
Article · September 2015
CITATIONS
6
READS
164
4 authors, including:
D. Ravikumar
DMI College of Engineering (Autonomous)
50 PUBLICATIONS233 CITATIONS
SEE PROFILE
Rubini Balakrishnan
Vels University
25 PUBLICATIONS46 CITATIONS
SEE PROFILE
Satheeskumaran Satheeskumaran
19 PUBLICATIONS759 CITATIONS
SEE PROFILE
All content following this page was uploaded by D. Ravikumar on 11 July 2020.
The user has requested enhancement of the downloaded file.
IJSART - Volume 1 Issue 9 –SEPTEMBER 2015 ISSN [ONLINE]: 2395-1052
Page | 43 www.ijsart.com
ASIC Implementation of an Adaptive Noise Canceller
for ECG Signal Processing Applications
K.Sasikala
1
, D.Ravikumar
2
, B. Rubini
3
, S.Satheeskumaran
4
1, 3
Department of EEE
2, 4
Department of ECE
1, 2, 3
School of Engineering, Vels University, Chennai, India
4
Karpagam College of Engineering, Coimbatore, India
Abstract- Electrocardiogram (ECG), a noninvasive technique
is used as a primary diagnostic tool for cardiovascular
diseases. A cleaned ECG signal provides necessary
information about the electrophysiology of the heart diseases.
It provides valuable information about the functional aspects
of the heart and cardiovascular system. The objective is to
automatic detection of cardiac arrhythmias in ECG signal.
This work focuses on developing a sophisticated, small and
reliable ASIC (Application Specific Integrated Circuit) chip
that can be used for monitoring and detecting the rate of heart
beat for heart transplantation patient. Well known adaptive
noise cancellation techniques are such as LMS (Least Mean
Square) and RLS (Recursive Least Mean Square) have been
extensively used for noise cancellation techniques with good
performance. The proposed architectures have been modeled
and verified for their functionality. Using the entire ASIC flow,
suitable results obtained at various stages are compared and
reported. The high computational requirement of all adaptive
filtering algorithms has limited the scope of its use in medical
applications. However, with rapid advances in VLSI (Very
Large Scale Integration) technology, it is possible to
implement complex circuits in a single chip. This work focuses
on developing architectures for adaptive noise cancellation
and its ASIC implementation.
Keywords- Heart rate variability (HRV), Adaptive filter, Fast Fourier
transform (FFT), Electrocardiogram (ECG)
I. INTRODUCTION
There is an enormous demand for reducing size and
power of transferable devices, used for monitoring critical
signals such as electrocardiogram (ECG),
electroencephalogram (EEG) and electromyogram (EMG).
Besides biomedical products, there are large number of
emerging healthcare applications that involve sensors and their
associated precise instrumentation and signal conditioning.
Low power, miniaturized and low cost monitoring/sensing
devices are the key components in such systems. The
performance of these devices directly depends on analog
signal conditioning, which must extract and amplify extremely
small signals from a noisy environment. Myopotential
spectrum is predominant at higher frequencies and
significantly overlaps with the spectrum of the ECG signal,
primarily with the spectrum of the QRS complex [1]. Thus,
the automatic interpretation, following accurate detection of
characteristic ECG points and waves, and measurement of
signal parameters, become difficult. EMG noise is caused by
increased muscle activity. The ECG signal is used to know the
cardiac condition of an ambulatory patient. Wireless
Ambulatory ECG recording is now routinely used to detect
arrhythmias and cardiac abnormalities. As the ECG signal
contains numerous artifacts, these artifacts have to be removed
before monitoring, from the receiver point-of-view, so that a
correct decision can be taken. So, it is necessary to remove the
different artifacts present in the ECG signal hence there is a
need of filtering the ECG signal. In a practical case most of
the signals are nonstationary and the filter, which we use must
change its coefficient according to the input signal. Several
filtering techniques have been presented in literature for ECG
analysis, which includes, adaptive and non adaptive
techniques [11]–[13], adaptive filtering techniques permit to
the detect time varying potentials and to track the dynamic
variations of the signals.
Electrical activity of heart can be recorded with
surface electrodes on chest or limbs. ECG wave shape may be
altered by cardiovascular diseases, atrial fibrillation, and
ventricular fibrillation and conduction problems. ECG signal
comprises of P wave, PG segment, QRS complex, ST segment
and T wave. QRS complex wave shape is affected by
conduction disorders. Ventricular enlargement could cause a
wider than normal QRS complex. The ST segment may be
depressed due to myocardial infarction. Presence of noise is
one of the most challenging problems in Signal Processing
basically due to the fact that a signal can pick up noise and be
distorted such that the information carried by the signal can be
misinterpreted. Thus, it is important that the impairments due
to noise is reduced or eliminated totally from signals in almost
all signal processing and communications tasks. Filtering is
widely used to remove the noise from the signal. However, in
the process, it also removes a part of the signal, which may be
an important part of the signal processing application.
The wavelet transform is an emerging signal
processing technique that can be used to represent real-life non
IJSART - Volume 1 Issue 9 –SEPTEMBER 2015 ISSN [ONLINE]: 2395-1052
Page | 44 www.ijsart.com
stationary signals with high efficiency [1]. Indeed, the wavelet
transform is gaining momentum to become an alternative tool
to traditional time-frequency representation techniques such as
the discrete Fourier transform and the discrete cosine
transform. By virtue of its multi-resolution representation
capability, the wavelet transform has been used effectively in
vital applications such as transient signal analysis [2],
numerical analysis [3], computer vision [4], and image
compression [5], among many other audiovisual applications.
Wavelet transform is mostly needed to be embedded in
consumer electronics, and thus a single chip hardware
implementation is more desirable than a multi-chip parallel
system implementation. However, time-varying autoregressive
models allow assessing, on a beat to beat basis, the spectral
parameters of HRV signal in a fast and efficient way
independently on the transitory events found through the
whole night recording (provoked by arousals, body
movements, and changes on sleep stages or apneas).
In the last few decades the demand for portable and
embedded digital signal processing (DSP) systems has
increased dramatically. Applications such as cell phones,
hearing aids, and digital audio devices are applications with
stringent constraints such as area, speed and power
consumption. These applications require an implementation
that meets these constraints with the shortest time to market.
The possible alternative implementations that can be used
range from an ASIC custom chip, general purpose processor
(GPP) to DSP processors. While the first choice could provide
the solution that meets all the hard constraints, it lacks the
flexibility that exists in the other two, and also its design cycle
is much longer. FPGAs prove particularly useful in data path
designs, where the regular structure of the array can be utilized
effectively. The programmability of FPGAs adds flexibility
not available in custom approaches, while retaining relatively
high system clock rates.
II. RELATED WORK AND ISSUES
The nonlinear filter that uses reversible WT allows
estimating noise level in individual decomposition bands and
proportionally adapting correction of WT coefficients. In this
way, we can achieve effective noise suppression while
distortion of the ECG signal is minimized. Besides the choice
of decomposition and reconstruction filter banks, the choice of
the level of decomposition and the strategy of WT coefficient
adjustment are also important. Different strategies of
thresholding the WT coefficients with down sampling are
discussed in [4]. In [5], the author attempts to optimize the
threshold parameters for a wavelet filter with WT with
decimation, and concludes that the optimal parameter values
depend on the level of interference. The disadvantage of
filtering with WT with down sampling is that the result is
dependent on the choice of the beginning of the filtering and
the need for interpolation in reverse transform, which is
always a source of errors. Transform without down sampling,
the so called stationary (redundant) wavelet transform (SWT),
is more preferable for filtering. Thresholding using SWT is
solved in [6]. Better results can be achieved by using the
wavelet Wiener filtering, when each transform coefficient is
adjusted separately. The Wiener filter requires an estimate of a
noise-free signal, which is necessary to calculate the
correction factor for the adjustment of transform coefficients.
The principle of the method was described in [7], where the
estimate of the noise-free signal was performed using another
wavelet filter, both implemented with decimation. The wavelet
Wiener filtering (WWF) with decimation and with simplified
estimation of the noise-free signal was used in [2]. In [8],
SWT with estimation of the noise-free signal was used. The
estimation was carried out with WT with decimation and hard
thresholding. In [9], both the transforms are stationary; the
estimation of a noise-free signal was carried out by
nonnegative garrote thresholding. The filters were tested on
signals with artificial noise, whose power spectrum was
adapted to the spectrum of an EMG signal. The parameters of
all the Wiener filters mentioned were set intuitively. The
authors of all the papers cited used dyadic transforms.
A flowchart demonstrating the signal processing
steps of the Pan and Tompkins algorithm (Pan & Tompkins,
1985) for the classical derivative-based QRS detection is
shown in Fig. 3. The ECG signal first passes through a set of
linear processes, including a band-pass filter comprising a
cascaded low-pass and high-pass, and a derivative function.
Non-linear transformation is then employed in form of a signal
amplitude squaring function. Finally, moving window
integration is performed before an adaptive threshold is
applied for detection of the QRS complexes. The underlining
principle of the algorithm is the detection of the slope of the R
wave through the derivative function, amplified by the
squaring function. The moving-window integration then
provides wave-form feature information in addition to the
detected R wave slope. Different from conventional method,
in our system, as we are only interested in the RR interval in
HRV analysis, we choose to assign an R peak to each detected
R slope from the output of the squaring function through an
adaptive threshold. Thus, we only require the band-pass filter,
derivative function, squaring function, and adaptive threshold
in our system. After differentiation, squaring function is
employed to enhance the characteristics of the signal. Then a
threshold is applied to the squared signal to detect the start of
the QRS complex. The peak of the squared signal is identified
as the R peak of the ECG data.
End of Document
203
Figures
