Trying to make spectogram for my IQ signal from FMCW radar

I am trying to make spectogram of my IQ signal from FMCW radar uRAD. I tried the sygnal analyzer.

Why it isn't like that

But just something like that

or that

It always like zero frequency and i dont know why because when i measure i was trying to do dynamic scenarios.

[Mathieu NOE]

hello

can you share some data ?

Answer 1

[Anumeha]

Hi Fabian,

Based on the graphs you have attached, it seems that the DC component is strong and as such shows up in the spectrogram.

The simplest way of removing it would be to subtract the signal by its mean.

Here is a simple program I wrote illustrating the same:

% Parameters 

fs = 48000;             % Sample frequency 
t = 0:1/fs:5;            % Time vector  
f0 = 1000;               % Initial frequency 
signal = chirp(t, f0, t(end), 10000) + 0.5*randn(size(t));  

% Spectrogram calculation 
window = 512;            % Window size 
noverlap = 256;          % Number of overlapping samples 
nfft = 2048;             % Number of FFT points 

figure; 
spectrogram(signal, window, noverlap, nfft, fs, 'yaxis'); 
xlabel('Time (seconds) in samples'); 
ylabel('Doppler frequency shift (Hz)'); 
colormap nebula;            
set(gca, 'YDir', 'normal'); 

Kindly refer to the documentation of “spectrogram” function linked below:

https://www.mathworks.com/help/signal/ref/spectrogram.html

Hope this helps!