Hello,

I'm reading the attached pdf file. I don't understand the design approach (yellow-highlighted). Could you please explain?

Thank You,
BlackMelon

 

Do you understand how to design a bandpass filter?

 

More specifically what is your knowledge of digital filtering?
The paper you referenced is a somewhat advanced discussion that presumes you are already familiar with the basics of digital filters, IIR and FIR types, for example.
Are you?

 

I am taking a Digital Signal Processing class. (Let's say it's ended cuz I'm waiting for my finals). I know formulas and steps to design those filters, but I don't know how to apply those formulas to this paper.

 

The paper states:
The structure of FIR filter can achieve a strict linear phase, but the computational amount does not meet the real-time system. The structure of IIR filter is simple to calculate, but it is very difficult to realize the strict linear phase.In this paper, the IIR structure is adopted to realize the simple integer coefficient digital filter of FIR, which can meet the requirements of realtime ECG signal processing and real-time performance. Design approach is to design a band-pass filter, and then design an all-pass filter which has the identical delay and gain with the band-pass filter. Finally, the all-pass filter removes simple integer coefficient filter to get the filter we are need.

My cursory understanding of this is that integer processing for the ECG filter is desired to simplify the processing requirements, so an IIR filter is utilized to get integer processing, along with an all-pass filer to get the desired linear phase response required from the filter.
So the paper describes how the IIR filter along with the all-pass function achieves these objectives.
You should be able to figure out the rest from there.

P.S. I suggest you don't use the word "cuz" in your finals.

 

The digital filter is essentially a weighted average.
For example you have the following numbers:
x1, x2, x3,x4,x5,..................They are your signal

You want to get the data:
y1,y2,y3,y4,y5,..........................Which is the filtered signal

Let's filter for example:
y(n)=(x(n-1)+x(n)+x(n+1))/3
That is, the new value will be the arithmetic mean between it and its neighbor.

Making the arithmetic mean we eliminate the sudden variations. So we made a low-pass filter.
It is interesting to see which filter in the electronics corresponds to and with what transfer function.

Executes digital filters in programming languages.Give numerical values and observe what operations do between input data.