• State Resolved
  • Locked Locked
  • Replies 15 replies
  • Answers 2 answers
  • Subscribers 31 subscribers
  • Views 949 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.

AWRL1432: Questions about cfar and BPM mode

J G1
Intellectual 835 points
Part Number: AWRL1432

Hi experts, 

   1.The cfar parameters (winLen, guardLen, thresholdScale...) that can be set in the current SDK are all range dimensions. Does Doppler dimension use the same parameters as range dimension? Can the parameters of range dimension and Doppler dimension be set separately?

    

    2.Currently, the SDK only supports BPM mode. Is Doppler compensation and phase compensation done in the SDK in this mode?

    3.When will the SDK supporting TDM mode be released?

    Thanks!

  • 0
    TI__Mastermind 40335 points

    Hi, there:

    Which demo you are looking at?

    Best,

    Zigang

  • 0 in reply to zigang Yang
    Intellectual 835 points

    Hi, 

        I'm using the mmw demo in SDK5.3. 

  • 0 in reply to J G1
    TI__Mastermind 40335 points

    HI, there:

    1) You can find the detail documentation about this demo at C:\ti\MMWAVE_L_SDK_05_03_00_02\docs\MmwaveDemo_documentation.pdf

    CFAR check is running only on the range domain at this time.  If you want to add the CFAR check also on the Doppler domain, you will need to add the code yourself. or wait for the future SDK release.  

    Can you also let me know what application you are trying to use radar for?

    Best,

    Zigang

  • 0 in reply to zigang Yang
    Intellectual 835 points

    Hi,

        We want to use it on a bicycle, because it is outdoors, which means the target speed may be very high (>100km/h).
      

  • 0 in reply to J G1
    TI__Mastermind 19520 points

    Hello.

    Did Zigang's latest response resolve your question?

    Sincerely,

    Santosh

  • 0 in reply to Santosh Krishnan
    Intellectual 835 points

    Hi Krishnan，

        I'm still waiting for responses to Q2 & Q3.

       

  • 0 in reply to J G1
    TI__Mastermind 19520 points

    Hello.

    J G1 said:
    Currently, the SDK only supports BPM mode. Is Doppler compensation and phase compensation done in the SDK in this mode?

    Phase compensation is done in this mode as it is part of the configuration.  Let me look into doppler compensation and see if it is done; I will follow up by Thursday with a response.

    J G1 said:
    When will the SDK supporting TDM mode be released?

    TDM mode support is available in the motion and presence detection lab, but I can look into this as well and follow up by Thursday with a response.

    Sincerely,

    Santosh

  • 0 in reply to Santosh Krishnan
    Intellectual 835 points

    Hi，

        To be clear, I am using the mmw demo, since the target speed I need to measure is fast (>100km/h), I need to confirm whether Doppler compensation is done in the SDK.

        In addition, since there is no velocity disambiguate in the SDK now, I plan to use gtrack to do velocity disambiguate, but if the point cloud speed is blurred, the doppler compensation is inaccurate, which means I can only use 1tx because In this case, doppler compensation is not required.

        Q1: Is my understanding correct?

        Q2: In order to only use 1tx, I tried setting TxChCtrlBitMask to 1, and then setting chirpTxMimoPatSel to 0. I found that it only prompted "Error: c_ChirpTxMimoPatSel must have value 4 (BPM-MIMO)", but the radar could still detect the target normally. Before doing more detailed testing, I would like to confirm that SDK5.3 is OK in this configuration?

        Thank you!

  • 0 in reply to J G1
    TI__Mastermind 19520 points

    Hello.

    J G1 said:
    Is my understanding correct?

    I have added this to my research of doppler compensation, and will follow up with you by Thursday on this.  I think that is correct, but let me confirm that just to be safe.

    J G1 said:
    2: In order to only use 1tx, I tried setting TxChCtrlBitMask to 1, and then setting chirpTxMimoPatSel to 0. I found that it only prompted "Error: c_ChirpTxMimoPatSel must have value 4 (BPM-MIMO)", but the radar could still detect the target normally. Before doing more detailed testing, I would like to confirm that SDK5.3 is OK in this configuration?

    If you get such an error during the CLI parsing, the device should not start, so this configuration is probably not correct.  Can you send me the configuration you are using so that I can recreate this on my end?  

    Sincerely,

    Santosh

  • 0 in reply to Santosh Krishnan
    Intellectual 835 points

    Hi,

        Here is the profile I use:

    1432_Profile.cfg

        Thank you!

  • 0 in reply to J G1
    TI__Mastermind 19520 points

    Hello.

    I believe this configuration works only because in the mmw_demo, BPM mode is hardcoded regardless of whatever is sent as part of the configuration.  So even if you configure it for TDM mode, it will still run as if it is in BPM mode, which is why it is working without error.  However, there are some other issues I noticed with your configuration/errors I saw with your configuration.  For example, you had a very large number of ADC samples which also threw an error.  If you want to tune your configuration, try using the sensing estimator tool linked here.

    Sincerely,

    Santosh

  • 0 in reply to Santosh Krishnan
    Intellectual 835 points

    Hi Santosh,

        1.I don't quite understand about "So even if you configure it for TDM mode, it will still run as if it is in BPM mode", could you please confirm what potential issues there might be in this case?

        2.Could you also confirm whether doppler compensation is implemented in the SDK?

    Thank you!

  • 0 in reply to J G1
    TI__Mastermind 19520 points
    J G1 said:
    I don't quite understand about "So even if you configure it for TDM mode, it will still run as if it is in BPM mode", could you please confirm what potential issues there might be in this case?

    There shouldn't be any issue because BPM mode is hardcoded into the processing chain for the mmw_demo in the SDK.

    J G1 said:
    2.Could you also confirm whether doppler compensation is implemented in the SDK?

    Doppler compensation is implemented in the SDK in the aoa2dproc DPU.

    Sincerely,

    Santosh

  • 0 in reply to Santosh Krishnan
    Intellectual 835 points

    Hi Santosh,

        Thank you for your reply! I have one last question to confirm.

        If I use 2tx and set it to TDM mode, I can understand that it is still executed in BPM mode, but in the 1432_Profile.cfg I sent before, only 1tx is used, and BPM need at least 2tx.

        As discussed before, I need to use 1tx to avoid problems with doppler compensation if the target speed exceeds the maximum unambiguous velocity:

    J G1 said:
    In addition, since there is no velocity disambiguate in the SDK now, I plan to use gtrack to do velocity disambiguate, but if the point cloud speed is blurred, the doppler compensation is inaccurate, which means I can only use 1tx because In this case, doppler compensation is not required.

        Also, I made a mistake in 1432_Profile.cfg. In order to use 1tx, I think I should set the value of ChirpTxMimoPatSel to 0.

        

        In this case, if BPM mode is hardcoded into the processing chain for the mmw_demo, I don't quite understand why the radar can correctly calculate the angle of the target, is there possibly some problem that I haven't discovered yet?

        The project I am working on now determines that I cannot wait for the new SDK, so I really need to make sure that there is no problem using 1tx in the mmw demo of SDK5.3.

        Thanks!

  • +1 in reply to J G1
    TI__Mastermind 19520 points

    Hello.

    As per the documentation, this demo can only support a 2 Tx configuration.  However, the number of channels is not hardcoded, so it looks like you need to make sure that the configuration enables 2 tx channels.  I believe you are still getting good phase info because you are using multiple Rx antennas, so the angle approximation and velocity approximation will be good.  As far as overall demo behavior, using the incorrect number of Tx may result in unexpected behavior as the demo only supports a 2-Tx configuration, but it is something I can look into and get back to you on.

    Sincerely,

    Santosh