DCA1000EVM: I want to create a 2d fft image through the captured raw data.

Hi,

Thanks for reaching out on our e2e forums! Can you please share what your image looks like in matlab?

Thanks,

Angie

Thank you for your apply

As I said, This image is the result of modifying the last part of the MATLAB example in swra581b(9.2).

Thanks,

Yeon Woo.

Hi,

The adcData will need to be reshaped per chirp and per number of antennas. An example of using the reshape function can be seen in the 9.2 example already. This function will need to be used to organize the ADC samples by antenna and by chirp.

I would recommend doing the fft in two steps, first do an fft per chirp / per antenna and then perform the second fft. 

Thanks,

Angie 

Maybe I don't quite understand what you're saying.

First, this part is added to the example 9.2. according to your reply.

%organize data per RX
adcData = zeros(numRX,numChirps*numADCSamples);
for row = 1:numRX
    for i = 1: numChirps
        adcData(row, (i-1)*numADCSamples+1:i*numADCSamples) = LVDS(i, (row-1)*numADCSamples+1:row*numADCSamples);
    end
end

% ADC data per frame
adcData = adcData(1, 1:numADCSamples*numChirps);
adcData = reshape(adcData, [numADCSamples, numChirps]);

% return receiver data
fft_out = fft2(adcData, numADCSamples, numChirps);
fft_out_test = fft_out;
fft_out = abs(fftshift(fft_out));
fft_out = 20*log10(fft_out);

%plot
figure(1)
imagesc(fft_out)

As a result, I was able to obtain the following image.

However, it is still different from the 2d fft image seen in the mmwave studio. 

So, I want to know 

Angie Mitchell said:

I would recommend doing the fft in two steps, first do an fft per chirp / per antenna and then perform the second fft

1. Is this different from the process of going through 2d fft?

2. And is there any github that I can refer?

Thanks,

YeonWoo

Hi YeonWoo,

Please keep in mind that the coloring in matlab may not be exactly the same as the coloring in studio. This looks like it may be correct but with both axes flipped.  Below is a code snippet that may help explain what I mean by "I would recommend doing the fft in two steps, first do an fft per chirp / per antenna and then perform the second fft"

Thank you,

Angie

I flipped the axis, but it's still different from what I saw in mmwavestudio. (Right one is an image from mmwavestudio.)

How can you make this the same as the picture above?

Hello,

I am looking into this and will get back to you soon.

Thanks,

Angie

Thank you for your sincere response. In the .m file you uploaded, variables such as the number of chirp and the number of frames were adjusted according to the situation. But I still can't see the result I want. In mmwave studio, I can see that the axis is speed and distance, but is this not relevant?

Hi Yeon,

I am sorry but it is out of my scope to explicitly help you with custom code on this forum. I have already stepped a little outside of the scope of e2e by giving you a .m script to try. 

Thank you,

Angie