ECG Live Monitoring/Visualization + Data conversion on Matlab - matlab

Question

I am working on ECG project for biometric identification: I use Arduino + Olimex ECG/EKG Shield and transmit the data to MatLab.
I am facing two issues:
Monitoring on real time: The X-Axis scale changes automatically and will compress the ECG after long time. I want to visualize a specific view so QRS complexes will be clearly visual. Any generic code?
Getting real voltage values: Analog.read() on Arduino converts voltages from 0-5 to 0..1023 and then we send them via serial communication. On Matlab, ECG does not have specific min and max to get the real values of the voltages to extract the characteristics. Any ideas about how that could be done.
Thank you so much in advance.
Waiting for your replies.
Kind regards,
Badreddine

Answer 1

edit disclaimer, I wrote this code from memory, so it may be buggy, apologies in advance
1.Lets say your data is stored in some array data_array and every time we get a sample we do something like this
data_array = [data_array, sample];
Now let's say in our viewer we only want to view 20 samples. we can do something like this
num_visible_samples = 20; %define this somewhere outside your data aquisition loop
%gets new data somehow
data_array = [data_array, new_data_point];
%plots all data
plot(data_array);
length_data_array = length(data_array);
if (length_data_array >= num_visible_samples)
%this makes the plot window only show the current data point and the
%previous few data points, the Yaxis will automatically be scaled
xlim([length_data_array -num_visible_samples, length_data_array ]);
end
another thing you could do (which will be faster) is to only plot the number of samples you desire
num_visible_samples = 20; %define this somewhere outside your data aquisition loop
%gets new data somehow
data_array = [data_array, new_data_point];
%plots all data
plot(data_array);
length_data_array = length(data_array);
if (length_data_array >= num_visible_samples)
%plots only the specific number of data points
plot(data_array(end-num_visible_samples,end));
%im not sure if you need to specify the axis, but i think you do
xlim([length_data_array -num_visible_samples, length_data_array ]);
else
plot(data_array);
end
2.Converting the values back and forth is pretty simple. This is jsut some psuedocode
maxV = 5.0;
minV = 0.0;
num_of_quantized_values = 1024;
%this says how much the lsb of the adc is, basically we just divide our
%input range (0-5 volts) by our output range (0-1023) => 4.88mV
adc_resolution = (maxV - minV) / num_of_quantized_values;
to convert the digital value back to an analog voltage we multiply our resolution by the value returned
val_in_volts = adc_digital_reading * adc_resolution;
so for example, say the adc read 256
256 * 4.88mV = 1.25
edit2
Since atmega chips usually have 10bit ADCs you get the data adc data in two bytes. The lower 8 and then the upper 2. The fix isn't too bad, say we have two variables
lower_byte is uint8
upper_byte is uint8
so I would do it this way (there are various methods depending on implementation)
adc_digital_reading = bitor(bitshift(uint16(upper_byte ),8), uint16(lower_byte));
%mirrors this c code
%adc_u16 = (upper_byte << 8) | lower_byte;
The important part is converting the uint8 types to uint16, or it wont work properly
to be really safe you could do
adc_digital_reading = bitand(bitor(bitshift(uint16(upper_byte ),8), uint16(lower_byte)),1023);
%adc_u16 = ((upper_byte << 8) | lower_byte) & 0x03FF;