• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 31 subscribers
  • Views 1089 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.

IWR1443: Angular resolution

allen dominek
Expert 1515 points
Part Number: IWR1443

Hi,

I thought I knew what was meant my angular resolution based upon a number mmwave documents I examined but after taking some measurements, I found that I need some clarification.  I ran the standard mmwave demo application using 2 transmitter and 4 receiver antennas except that I increased the cfar threshold to 4000.  I collected the detected object data when a dihedral moved across the the field of view and back again.  I was  expecting to observe a "digitized" variation in x since the resolution is suppose to  be 15 degrees.  Instead, I observed the following:

1. The typical frame had a number of detected objects even though the cfar threshold was set to 4000 (I thought such a threshold would have eliminated many objects that had a smaller return magnitude).  For frame 16, I observed

       Frame   Range   Velocity   PeakVal     X                 Y
                      Index     Index       Index

          16          82      1         142    -0.523437500    3.31250000    
          16          86      1        4281   -0.439453125    3.49023438    
          16         185      1          47    -0.708984375    7.53515625    
          16         169      2          45    -0.648437500    6.88281250    
          16          88      7          46     -1.68750000       3.17968750    
          16          88     -7          39     1.23828125        3.38085938    
          16         180     -1          67   -0.689453125     7.33007812    
          16          82      1          82    -1.67773438        2.90429688    
          16          86      1        2174   1.53906250        3.16406250    
          16         169      2          28    3.24023438        6.10742188    
          16          88     -7          20     3.26171875        1.52148438

2. When I selected the X and Y values for the detected objects based upon the largest PeakVal, I saw a very smooth variation in x as shown in the attached image.

a. Can you tell me why the x variation was so smooth when I thought the angular resolution was 15 degrees?  The field of view was +/- 25 degrees.

b. Can you  tell me why I collected so many detected objects when the cfar threshold was set to a fairly large value of 4000.

c. I was surprised that there were two fairly large returns (4281 and 2174).  The 2nd largest return had a x value that was positive (the true value is negative) and a y value twice of what it should have been.  There was only one dominant scatterer (dihedral) in the scene.  Any thoughts?

d.  Would there be any settings to filter out some of the erroneous detected objects?

Thanks in advance.

Al

The configuration used was

sensorStop
flushCfg
dfeDataOutputMode 1
channelCfg 15 5 0
adcCfg 2 1
adcbufCfg 0 1 0 1
profileCfg 0 77 429 7 57.14 0 0 70 1 240 4884 0 0 30
chirpCfg 0 0 0 0 0 0 0 1
chirpCfg 1 1 0 0 0 0 0 4
frameCfg 0 1 16 0 100 1 0
lowPower 0 0
guiMonitor 1 1 0 0 0 1
cfarCfg 0 2 8 4 3 0 4000
peakGrouping 1 1 1 1 229
multiObjBeamForming 1 0.5
clutterRemoval 1
calibDcRangeSig 0 -5 8 256
compRangeBiasAndRxChanPhase 0.0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0
measureRangeBiasAndRxChanPhase 0 1.5 0.2
CQRxSatMonitor 0 3 5 123 0
CQSigImgMonitor 0 119 4
analogMonitor 1 1
sensorStart

  • 0
    TI__Guru 50240 points
    Hi Allen,


    a. Can you tell me why the x variation was so smooth when I thought the angular resolution was 15 degrees? The field of view was +/- 25 degrees.

    So just to clarify, Angular Resolution (similar to range resolution) is the difference in angle required for the radar sensor to distinguish two different objects. It IS NOT a measure of accuracy.

    b. Can you tell me why I collected so many detected objects when the cfar threshold was set to a fairly large value of 4000.

    I really cannot make a statement on this because I have not seen the scene you are trying to evaluate. I can comment a little if you send a picture but it appears that all of the detected objects have sufficiently passed the threshold for detection. Perhaps raise the CFAR to 4200, 4500, and then 5000 and then re-evaluate. It is is interesting since you have two objects with very high return values while the others are so much smaller.

    One other piece of information that might be helpful is to include doppler information. If some of these objects with lesser peakVal have a non-zero doppler then they might be false detections and your CFAR threshold isn't large enough.


    c. I was surprised that there were two fairly large returns (4281 and 2174). The 2nd largest return had a x value that was positive (the true value is negative) and a y value twice of what it should have been. There was only one dominant scatterer (dihedral) in the scene. Any thoughts?

    As per my above comment, I'd recommend comparing doppler information and then also adjusting CFAR threshold. It is very difficult to make other recommendations without any knowledge of your scene.

    d. Would there be any settings to filter out some of the erroneous detected objects?
    That setting is of course the Constant False Alarm Rate (CFAR).


    Cheers,
    Akash
  • 0 in reply to Akash Gondalia
    Expert 1515 points

    Hi Akash,

    Thank you for your response.  I would like to clarify my first question.

    I attached an image showing the x variation from the detected objects algorithm.  It was for a corner reflection moving a distance of ~3m.  The total field of view was -25 degrees to +25 degrees of view.  I had expected to get a curve with poor positional resolution since I thought the angular resolution was 15 degrees, so 50/15=~3.  The curve I showed was very smooth.  What is the range of the x,y coordinates in the raw output of the demo program?  What controls the resolution?  I thought is was the number of virtural receiving antennas.

    Thanks.

    Al