IWR1642BOOST: Capturing data in complex format without DCA1000

Hello,

1. The only way to get raw data over UART is using the DCA1000

2. I will get back to you with an answer on this tomorrow

Thank you, 

Angie

Hello Angie,

Thank you for your answer, I'm waiting for the second...

Have a nice day.

Hi Abdellah,

What type of information are you hoping to display on your time-frequency spectrogram? What do you want to be able to tell from the spectrogram?

Thank you,

Angie

If you are meaning the range-doppler data from UART, you would not need complex data. Doppler is already computed so you would just need to map the range-doppler into doppler if you are summing across range. 

Thanks,

Angie

Hi Angie,

Sorry i didn't specified the type of info to be plotted, in fact i need to plot the doppler info. (see figure)

I need this plot for getting movement signature, for further process. (on figure : walk signature)

Thank you for your help.

No i don't mean the range doppler plot because i already plot it and i found it insufficient for my work.

However, i would know if it's possible to pass from range-doppler to time-frequency (doppler) plot.

I hope i was clear.

Thank you again.

Hi Angie,

Thank you so much for your answers, i feel it's going in the good direction for resolving my issue,

Since i'm still beginner on TI cards, Could you please explain further for me how to modify the code (where to modify exactly) to get the doppler cross range output on the UART. (would i modify on the SDK ?)

By the way, i don't need the visualizer, because i'm reading the data over UART using my own Python program, and i'm plotting my own graphics also. so i need just the good data for it.

Greatings,

Abdellah.

Hi Muhammet,

Thank you for your answer it's helping me having some time-doppler plot, but i still confused because some resources says that we have to apply an STFT (not summing) on the range doppler matrix (article : Radar sensing for healthcare - Enlighten: Publications (gla.ac.uk))

here is an extract :

Thank you again and Angie too.

Have a nice day.

Hello Abdellah.

In the paper you shared, it is not clear how the authors related the range-Doppler with time-Doppler through STFT. They did not address how they handled the range domain or how they related the Doppler FFT with STFT (overlap factor, etc.).

In theory, STFT is computed using a sliding window over time (with a defined overlap factor) and performs a discrete Fourier transform on the windowed data. It assumes that the chirps are captured continuously. However, our sensors (as a typical automotive radar sensor) utilize a framing scheme (chirp blocks are sent frame by frame). Hence, an STFT is not directly applicable.

Here, we perform the Fourier transform within a frame window (i.e., Doppler FFT) to mimic the STFT with zero overlap factor across frames.

Regards.

Muhammet 

Thank you so much Muhammet for your answer it's really helping,

I'll have only one last question. Could you tell me how to do reverse engineering on the outputed data on UART for having the raw Data, i tried to apply the inverse 2D fft but it doesn't look like the original data.

I think that it's a processed data (after 2d fft application) we recieve on UART (see the below fig from official doc)

Thank's Muhammet for your answers, i really appreciate your help.

Have a nice day.

Thank you Abdellah for reaching out on our e2e forums and please feel free to make another thread if you have any more questions!

Angie