• State Resolved
  • Locked Locked
  • Replies 14 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 3230 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.

AWR1642BOOST: The false detection with AWR1642BOOST

Nomo
Guru 16800 points
Part Number: AWR1642BOOST
Other Parts Discussed in Thread: AWR1642

Hello,

Please let me know the cause and solution of the false detection with AWR1642BOOST.

In case the reflector is located at 1.5m far from AWR1642BOOST, the false detection is at the maximum detectable distance - 1.5m.
The maximum detectable distance is calculated by the following formula using profileCfg parameters.

The maximum detectable distance = (sampling rate) x c / 2 / slope = (5583e+3)*(3e+8)/2/(21.53e+12)=38.9m

The red circle in the following figure shows the false detection.
The false detection point is about 37.4m (38.9m - 1.5m).

The conditions are following.
-AWR1642BOOST is used for the evaluation.
-The firmware is pre-build binary in the mmWave SDK (version is v01.00.00.05).
-Demo visualizer is used for the evaluation and the version is v1.2.0.
-The configure file is attached.

Best Regards,
Nomo

sdkv1r1_SRR_180320.cfg

  • 0
    TI__Guru*** 149044 points
    HI,

    Please ask the customer to update firmware to mmWave SDK 1.2

    Thank you
    Cesar
  • 0 in reply to Cesar
    Guru 16800 points

    Hello Cesar-san,

    Thank you for your reply.
    However, they've already started their development with v01.00.00.05 of SDK.
    Could you tell me the reason and solution for the version?

    Best Regards,
    Nomo

  • 0 in reply to Nomo
    TI__Mastermind 24955 points
    Nomo,

    If you are using SDK version 1.0.0.5, then you need to make sure you select SDK version 1.0 in the online mmWave Visualizer. The SDK version you flash to your device needs to match the SDK version you select in the visualizer. See if changing the SDK version in the visualizer resolves this issue.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    Guru 16800 points

    Hello Kyle-san,

    Thank you for your reply.
    My customers have tried your advice; however, the false detection isn't resolved.

    Could you investigate this false detection, again?

    Best Regards,
    Nomo

  • 0 in reply to Kyle Cousino
    Guru 16800 points
    Hello Kyle-san,

    Do you have any update on my thread?
    My customers think that the factor of the fault detection is a image frequency.

    The image is filtered at ADC output; however, the image is not rejected completely.
    The signal is about 105dB and the fault detection (image) is about 64dB which is the result of filter.

    Is there any method to resolve the fault detection?
    They want to know the way to correct of IQ Imbalance.

    Best Regards,
    Nomo
  • 0 in reply to Nomo
    TI__Mastermind 24955 points
    Nomo-san,

    After looking at your plot again, it also looks like the radar is detecting a point at around 40 m at an even stronger level than the detection point you originally asked about. Are they any potential objects in the field of view of the radar that could be causing these reflections? In the meantime I will take a look into your question about correct IQ imbalances. Please let me know if you have any questions.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    Guru 16800 points
    Hello Kyle-san,

    Thank you for your reply.
    There are not any potential objects around 37.5 and 40m.
    Do you have any updated information?

    Best Regards,
    Nomo
  • 0 in reply to Nomo
    TI__Mastermind 24955 points
    Nomo-san,

    This could be related to the CFAR threshold. I would suggest adjusting the CFAR threshold and seeing how that impacts the outputs in your range plot.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    TI__Mastermind 24955 points
    Nomo-san,

    I have discussed this thread with our systems team, and they were able to provide some insight into what you are seeing.

    It is not uncommon for power from the DC FFT bin to wrap around to the last FFT bin. This can result in potential detections at the end of your calculated max range based on your profile configuration. To avoid this issue, the systems team recommends discarding the last 10% of the FFT range bins to avoid seeing false detections.

    Please let me know if you have any additional questions.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    Guru 16800 points

    Hello Kyle-san,

    Thank you for your reply.
    My customers have additional two questions; because, they want to know why the last 10% of the FFT range bins should be discarded to avoid seeing false detections.

    1.What is the cause of the 10% limitation?
      Is this the limitation of the filter's performance in complex 1x mode?

    2.In case of AWR1642, is there any method to avoid these fault detections by hardware?

    Best Regards,
    Nomo

  • 0 in reply to Nomo
    TI__Mastermind 24955 points
    Nomo-san,

    When implementing your desired chirp configuration, you should take into consideration the maximum range that you want to resolve. So for example, if you want to resolve a maximum of 20 meters, then it would be best to design a chirp with a maximum range of 22 meters. This takes into account the 10% cutoff mentioned in my earlier post.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    Guru 16800 points
    Hello Kyle-san,

    Thank you for your reply and I can understand what your suggestion.
    However, I want to know the reason why this solution is needed for AWR1642.

    Best Regards,
    Nomo
  • 0 in reply to Nomo
    TI__Mastermind 24955 points
    Nomo-san,

    As I previously stated in one of my posts, these reflections are created when taking the range FFT on the chirp data. That is why these reflections occur. Like in my previous post, it is not uncommon for power from the DC FFT bin to wrap around to the last FFT bin. This can result in potential detections at the end of your calculated max range based on your profile configuration. I have discussed this thread with our systems team to verify this behavior. That is why the systems recommends discarding the last 10% of FFT range bins to account for the reality of seeing wrapped around false detections. This phenomena would not be unique to AWR1642. It is merely the unavoidable reality of the behavior seen when taking the range FFT.

    Since you designed your original chirp for a maximum range of roughly 40 meters, implementing a 10% cutoff would remove the last 4 meters of FFT range bin data. This would remove both of the false reflections that you seen in your range plot.

    I am considering this thread to be closed now. Please create a new thread if you have any other questions.

    Regards,
    Kyle
  • 0 in reply to Kyle Cousino
    Guru 16800 points

    Hello Kyle-san,

    Thank you for your reply and I appreciate your kind support.
    I'll tell my customers your answer and I close this thread.
    In case additional questions occur, I will make another post.

    Best Regards,
    Nomo