• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 31 subscribers
  • Views 633 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWR6843ISK: Azimuth Static Heatmap

user5996765
Prodigy 30 points
Part Number: IWR6843ISK

Hi, 

According to the mmWave SDK, the Azimuth Static Heatmap is saved in the type of cmplx16ImRe_t.

And the document goes on to show that cmplx16ImRe_t is saved in int16_t.

My question is shouldn't the imaginary and real values of the heatmap a type float?

Thank you

Sincerely, 

Jonathan Tang

  • 0
    TI__Mastermind 37805 points

    Hi Jonathan,

    Why do you think the real/imag values need to be a float? They may be int16's saved in a Q7/Q9 format, or something similar to allow for decimal values while preserving the uint16_t type size.

    Best,

    Nate

  • 0 in reply to Nathan Block
    Prodigy 30 points

    Hi Nate,

    Thank you for pointing out the use of Q format since it has higher efficiency for DSP calculations.

    If that is the case what is the exact Q format for the Azimuth Static Heatmap?

    I also looked into the SDK documents, is the length of the data in 16 bit due to the size of the ADC?

    Thank you 

    Sincerely,

    Jonathan Tang

  • +1 in reply to user5996765
    TI__Expert 4448 points

    Hi Jonathan, I am taking over from Nathan here. 

    As Nathan said, I would question your assumption about why the heatmap value would be a float. That being said, on closer investigation, I do not believe that this is being stored in Q-format. 

    The output result of the azimuth static heatmap processing is a 2D FFT array in the range direction (cmplx16ImRe_t x[numRangeBins][numVirtualAntAzim]), at doppler index 0, used for azimuth heatmap display. The input is the 1D-FFT radar cube matrix.
    I would recommend you run a demo and parse out some responses to see if they make sense to you.
    The real and imaginary components are both 16 bits because this is what our C674x DSP requires for complex multiplication instructions.
    Regards,
    Luke