Prev: Figures opening as logical variables
Next: MSK (or OQPSK) delay problem
From: Jason on 12 Jan 2010 11:53
Hello all,
I am running into a small problem here with Matlab. I have a radar waveform that contains doppler content, and I am trying to plot it in a spectogram.
I am able to plot it on to a spectogram using the >spectogram command, but what I really wanted to do is to change the doppler frequency of the spectogram into the velocity such as m/s on the spectogram?
I am thinking that if I can get the frequencies of the spectogram, then I can just scale it into velocity, because I know the radar frequency and its wavelength.
But how excatly can I do that to get the frequency and then the velocity??? Please help??
From: Rune Allnor on 12 Jan 2010 11:59
On 12 Jan, 17:53, "Jason " <sy...(a)osu.edu> wrote:
> Hello all,
> I am running into a small problem here with Matlab. I have a radar waveform that contains doppler content, and I am trying to plot it in a spectogram.
> I am able to plot it on to a spectogram using the >spectogram command, but what I really wanted to do is to change the doppler frequency of the spectogram into the velocity such as m/s on the spectogram?
> I am thinking that if I can get the frequencies of the spectogram, then I can just scale it into velocity, because I know the radar frequency and its wavelength.
> But how excatly can I do that to get the frequency and then the velocity???

It depends entirely on the waveform. If the waveform contains
a chirp with one easily identifiable frequency refernece point,
identify that point in the spectrogram, and measure the Doppler
offset.

If you can't find such a reference point, you are in trouble.

Rune
From: Lothar Schmidt on 12 Jan 2010 12:03
Am 12.01.2010 17:53, schrieb Jason:
> Hello all,
> I am running into a small problem here with Matlab. I have a radar
> waveform that contains doppler content, and I am trying to plot it in a
> spectogram.
> I am able to plot it on to a spectogram using the >spectogram command,
> but what I really wanted to do is to change the doppler frequency of the
> spectogram into the velocity such as m/s on the spectogram?
> I am thinking that if I can get the frequencies of the spectogram, then
> I can just scale it into velocity, because I know the radar frequency
> and its wavelength.
> But how excatly can I do that to get the frequency and then the
> velocity??? Please help??

I would start thinking about the causes of doppler shift and try to find
books or articles in the internet with formulas describing this effect.
Then I would try to understand the formulas and even if I would bve too
stupid, I would probably try to write anything in matlab code
approximately resembling the formulas.
Then, if the program would not show the expected behaviour (what also
might be expected), I would ask for help here.
I'm shure that if you show your thoughts and work you will find someone
who corrects your wrong formulas and anotherone who corrects your code.

Well in the time that i wrote this answer i also could have tried to
solve your problem... sorry that I did not.
From: Jason on 12 Jan 2010 12:15
Rune Allnor <allnor(a)tele.ntnu.no> wrote in message <0f3ee851-2c85-47a7-884c-f8f88d565f66(a)k19g2000yqc.googlegroups.com>...
> On 12 Jan, 17:53, "Jason " <sy...(a)osu.edu> wrote:
> > Hello all,
> > I am running into a small problem here with Matlab. I have a radar waveform that contains doppler content, and I am trying to plot it in a spectogram.
> > I am able to plot it on to a spectogram using the >spectogram command, but what I really wanted to do is to change the doppler frequency of the spectogram into the velocity such as m/s on the spectogram?
> > I am thinking that if I can get the frequencies of the spectogram, then I can just scale it into velocity, because I know the radar frequency and its wavelength.
> > But how excatly can I do that to get the frequency and then the velocity???
>
> It depends entirely on the waveform. If the waveform contains
> a chirp with one easily identifiable frequency refernece point,
> identify that point in the spectrogram, and measure the Doppler
> offset.
>
> If you can't find such a reference point, you are in trouble.
>
> Rune

Thank you Rune,
What I am trying to do is to plot a velocity vs. time spectorgram similiar to the regular spectogram of frequency vs. time.
So I think what I am doing is basically try to readjust the frequency content that i got from the 'stft' of the spectogram and scale it using the formula
v = 1/2*(Fd/Fc)*c
where
v is the relative velocity of the moving object
Fd is the doppler frequency
Fc is the transmitted frequncy
c is the speed of the RF wave
The trouble I am having is knowing where can I get the frequency content and how to write a code to get it?

My code is this:
>Win=256;
>Nlap=Win-1;
>S=spectrogram(A, Win, Nlap, 'yaxis'); %A is my waveform
>[m,n]=size(S);
>imagesc(20.*log10(abs((S(1:m,:)))));
>shading flat;
>axis xy; %cartesian

I am guess the S i get from spectrogram is the fft complex values of the windowing across the entire waveform A, but I am confuesd in how to extract the frequency content along with time, so that i can plot it as a spectrogram of velocity vs time rather than frequency vs time??
Thanks a lot! I really appreicated the response.
From: Jason on 12 Jan 2010 12:25
"Jason " <syu.3(a)osu.edu> wrote in message <hii9dg$g2b$1(a)fred.mathworks.com>...
> Hello all,
> I am running into a small problem here with Matlab. I have a radar waveform that contains doppler content, and I am trying to plot it in a spectogram.
> I am able to plot it on to a spectogram using the >spectogram command, but what I really wanted to do is to change the doppler frequency of the spectogram into the velocity such as m/s on the spectogram?
> I am thinking that if I can get the frequencies of the spectogram, then I can just scale it into velocity, because I know the radar frequency and its wavelength.
> But how excatly can I do that to get the frequency and then the velocity??? Please help??

Thanks for the responses.
I am trying to plot a spectrogram that shows velocity vs time instead of frequency vs time.
So I was hoping to change the doppler frequency content of the spectrogram to velocity using the following formula:
v = (1/2)*(Fd/Fc)*c
v = velocity of moving object
Fd = doppler freq
Fc = transmitted freq of radar
c = RF wave speed of light

My code:
S=spectrogram(A, Win, Nlap, 'yaxis');
[m,n]=size(S);
imagesc(20.*log10(abs((S(1:m,:)))));
shading flat;
axis xy; %cartesian

and my waveform is in A, its a digitally sampled signal at 22KHz.

Is there any way to extract that frequency from S, which I am guess is the complex number resulted from stft of A? But exactly how to work it?
Thanks
 |  Next  |  Last
Pages: 1 2
Prev: Figures opening as logical variables
Next: MSK (or OQPSK) delay problem