• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 424 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.

IWR1642BOOST: Saving Relative Power as Opposed to SNR

Isaac Mears
Intellectual 530 points
Part Number: IWR1642BOOST
Other Parts Discussed in Thread: IWR1642

We are looking at the point cloud data given from the IWR1642BOOST. We noticed that one of the points in this point cloud is SNR, but we are more interested in looking at the relative power or intensity of the returned point. Is there a way to obtain this information?

Furthermore, if we knew how you guys calculate the SNR, could we do a reverse calculation to figure out the intensity based on this SNR value?

Thank you.

  • 0
    TI__Mastermind 30405 points
    Hello,

    Could you please check that you are looking at the documentation in the correct SDK? The latest SDK for IWR1642 is mmWave SDK 2.1 and as defined in the following doxygen document, the point cloud returned by SDK 2.1 OOB demo contains peakval i.e. the absolute intensity of the point from the range FFT spectrum:

    file:///C:/ti/mmwave_sdk_02_01_00_04/packages/ti/demo/xwr16xx/mmw/docs/doxygen/html/struct_mmw_demo__detected_obj__t.html

    The change to report SNR (instead of peakval) has been made in SDK 3.1 which does not have an OOB demo for IWR1642. SDK 3.1 contains OOB demos only for xWR68xx and xWR18xx devices.

    Regards
    -Nitin
  • 0 in reply to Nitin Sakhuja
    Intellectual 530 points

    I am using the People Counting lab in the industrial toolbox.

  • 0 in reply to Isaac Mears
    Intellectual 530 points
    Does this also give intensity instead of SNR when it outputs data to the IWR1642BOOST?
  • 0 in reply to Isaac Mears
    TI__Mastermind 30405 points
    Per the lab documentation, the Point cloud TLV provides Range, Angle, Doppler and SNR associated with each point. I have assigned this ticket to our people counting applications expert to respond further.

    Thanks
    -Nitin
  • 0 in reply to Nitin Sakhuja
    TI__Mastermind 28055 points
    Hi Isaac,

    With the data available from the People Counting demo, you will not be able to calculate peak value. We calculate SNR as the Peak Value/Noise - the demo does not return noise. I would like to understand what you are wanting to do with the peak value to see if there is another way for you to achieve your goal.

    Regards,
    Justin
  • 0 in reply to Justin Curewitz
    Intellectual 530 points

    We are trying to do tests to see how a person holding different types of objects can affect their peak value return. We then want to be able to take data of people walking around and create estimates for whether or not they are carrying those objects.

    We just completed testing using the demo visualizer to create peak value signatures for different objects, but the demo visualizer is very bad at tracking people because:

    a) It gives less returned points for the point cloud

    b) Points with nonzero Doppler are not given peak values

    What we want are the four values in the point cloud (range, angle, Doppler, peak value) at the resolution in the people counting lab so that we are able to both track people and characterize their movements and the things they are holding.

  • 0 in reply to Isaac Mears
    TI__Guru 50240 points
    Hi Isaac,

    The People Counting Lab is able to provide object data pertaining to the four components you listed: range, angle, Doppler, peak value.

    Have you tried debugging the People Counting Lab with Matlab in order to see how these values are changing?

    It is also possible to save a sequence of frames using the People Counting Lab run-time executable, so there are multiple ways to evaluate what you are trying to determine.

    Let me know if you have more questions.


    Cheers,
    Akash