• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 256 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.

IWR6843AOPEVM: CFAR

Gmsy Bryant
Expert 1650 points
Part Number: IWR6843AOPEVM
Other Parts Discussed in Thread: IWR6843

In the IWR6843 people counting demo, why should cells be discarded in the beam dimension when CFAR detection is performed？

  • 0
    TI__Mastermind 37805 points

    Hi,

    Power from a single reflection often ends up in a few range/velocity/angle bins due to the quantization of the FFT. As a result, guard cells are used to ensure that you're not comparing the object's power with the power in the cells right next to it, which will have some of the object's power smeared into them as well.

    Constant false alarm rate - Wikipedia

    Best,

    Nate

  • 0 in reply to Nathan Block
    Expert 1650 points

    Thank you very much for your answer. Maybe I didn't describe my question clearly. What I want to ask is the role of "discarded cells", not the role of “guard cells”. For an RA spectrogram, discarded cells are generally cells with several columns on the left and right sides and rows on the top and bottom sides. They do not participate in cfar detection as UCTs. I do not understand why there is such a strategy.Take a look at Figure 13 on page 52 of the following link.

    dev.ti.com/.../node

  • 0 in reply to Gmsy Bryant
    TI__Mastermind 37805 points

    Hi,

    Thank you for clarifying. The discarded cells are there to avoid unexpected spikes in the heatmap. Sometimes noise at extreme angles can have a higher variance due to the angle FFT and antenna pattern, so it's sometimes helpful to discard cells on the extreme edges to avoid false positives, or bad noise estimates for nearby cells.

    Best,

    Nate

  • 0 in reply to Nathan Block
    Expert 1650 points

     Why noise at extreme angles can have a higher variance due to the angle FFT and antenna pattern?

    If capon or other methods are used instead of FFT method when calculating angular spectrum, will similar problems occur

  • +1 in reply to Gmsy Bryant
    TI__Mastermind 37805 points

    Hi,

    The antenna pattern's directivity means that the received signal at different angles can vary greatly even if the object generating the reflection that creates the signal is the same. e.g.  an object that yields 20 dBm of reflected power at an angle of 0 degrees may yield 10 or 0 dBm at greater angles off the antenna's boresight. You are more than welcome to remove the discarded cells in the algorithm though if you want. This is in the CFAR code.

    Best,

    Nate

  • 0 in reply to Nathan Block
    Expert 1650 points

    OK, your answer is very helpful to me. Thank you very much.