• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 2 answers
  • Subscribers 34 subscribers
  • Views 1340 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: Wrong angle estimation on right side of AWR1642BOOST

Chidvilas Yerramsetti
Intellectual 480 points
Part Number: AWR1642BOOST

We're using AWR1642BOOST EVM for object detection(reflector) at 11m with the profile configuration:

profileCfg 0 77 7 5.8 83.5 0 0 39.2 1 256 3340 0 0 30

The result of DOA algorithm is found that when the reflector moves beyond 60deg on the right side it is being shown at 60deg on the left.

This issue is not seen when the reflector from 60deg on the left and beyond.

But the angle of detection for the reflector is correct for 0 to 60deg.

Could you please explain why this issue is specifically seen on the right side.

Thanks and regards,
Chidvilas

  • 0
    TI__Guru*** 147964 points
    Hi,

    Please let us know which SW version is being used.

    Is this behavior dependent on the range?

    thank you
    cesar
  • 0 in reply to Cesar
    Expert 1765 points

    Hi Cesar,

    We are using SDK 1.2 version.

    And yes, the phenomenon observed is related to the range.

    At 11m range, with corner reflector moving in an arc, the DoA angle seems to be correct till 60deg. At 5-6m range, the DoA seems to be correct till 80deg.

    Hope this information helps.

    Expecting at least pointers to help resolve this situation soon.

  • 0 in reply to Santhana Raj
    TI__Guru 69960 points

    Hello Santhana,

    If you go over the MIMO application note which explains the angle and FOV based on antenna design.

    As on the EVM inter-antenna distance is fixed and based on lamda (radar RF frequency, which varies from 76Ghz to 81Ghz) the maximum-FOV would be lesser than 90 degrees. For e.g if lambda is 3.95 mm (corresponding to 76 Ghz), and d is 3.8mm (as on EVM), then the maximum FoV @ 76 Ghz is 74 degrees. Any target beyond this angle will wrap around and appear at -74 degrees.

    Also, the variance of \phi depends on the SNR of the target. If the target has high SNR (for e.g. when the object is closer), the estimate of  \phi will be stable to a larger \phi than if the target has low SNR. Could you try the same experiment with a larger corner reflector? We've noted that errors in estimation increase at these extreme angles. 

    Regards,

    Jitendra

  • 0 in reply to Jitendra Gupta
    Expert 1765 points

    Hi Jitendra,

    Thank you for your detailed and patient explanation.

    The relation to the higher RCS object and the less than 90 degrees is as expected as you have explained from the theoretical basis.

    Our question is that, we see this phenomenon on only one side (Positive angles) and is not seen on negative angles at all. Sorry if we haven't been clear on this earlier.

    Your explanation, though is completely acceptable, doesn't explain why it will be partial to one side alone.!!

    Thanks in advance,

    Santhana Raj

  • 0 in reply to Santhana Raj
    TI__Expert 5235 points
    Not sure if this could be the problem but checking: I know you mentioned you see the detection corresponding to 0 to 60 deg but when you set your target at true 0 deg say, do you actually detect it at 0 deg or does it have an error (skewed on one side)? If measured is not matching the actual, then you may need to do calibration (run the range bias and phase calibration procedure in the demo) to get back symmetry.
  • 0 in reply to Piyush_
    Expert 1765 points
    Hi Piyush,

    I cam across this earlier and we too suspected the same. In our testing we see about 5-8 deg shift in measurement. We are not planning to do any calibration as of now.
    But this 8 degree shift shouldn't cause a variation for angel values from 60 deg to 90deg. The 8 deg variation at the centre seems to very to 30 deg shift at the sides!!
  • 0 in reply to Jitendra Gupta
    Prodigy 40 points

    Hi Jitendra,

    Thank you for your detailed explanation. 

    I have a small doubt on the calculation mentioned above though. You mentioned that the d is 3.8mm on the EVM. 

    Is this value mentioned in any document?. Also is this value 2d?

    Thanks and Regards

    Avinash K

  • 0 in reply to Avinash Kondeppady
    TI__Guru*** 147964 points
    Hi,

    I will try to reproduce this issue on our setup. Please let me know if the following is correct:

    1) Use SDK 1.2
    2) Update profile included in SDK with following - Which profile file did you update?
    profileCfg 0 77 7 5.8 83.5 0 0 39.2 1 256 3340 0 0 30

    3) Place a corner reflector at range ~10m, angle 0

    4) Move to the left ~90deg - observe detected point on GUI

    5)Move to the right ~90deg - observe detected point on GUI

    Is this correct?

    Thank you
    Cesar
  • 0 in reply to Cesar
    TI__Guru*** 147964 points
    Hi,

    I have reviewed the thread information and I would like to mention that for the AWR1642BOOST EVM the horizontal 6-dB-beamwidth is approximately +/- 50degrees (6-dB drop in the gain as compared to bore sight).

    So, we have not validated the behavior at angles 60-90degrees.

    It is possible that the behavior noticed is due to the wrapping around. For 60-90 angles the behavior is indeterminate.

    Thank you
    Cesar
  • 0 in reply to Cesar
    Intellectual 480 points

    Hi,

    From the antenna radiation pattern in the horizontal direction, it found that it is not symmetrical. There is about 1db difference between the right and left sides.

    This might be the reason for this behavior.

    Thanks,

    Chidvilas