• State
  • Locked Locked
  • Replies 8 replies
  • Subscribers 32 subscribers
  • Views 329 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.

AWR6843: Code is crashing at AoAProcHWA_peakGrouping() function

V Prathith Vasudev
Prodigy 70 points
Part Number: AWR6843

Hi Team,

I and my team have been experimenting with changing profile configuration in the demo for xwr64xx in mmwave sdk version 3.5.0.4. Also we have changed the antenna geometry to as follows:

     .txAnt = {
                {0, 0},    
                {0, 1},    
                {0, 2}
       },
     .rxAnt = {
                {0, 0},   
                {1, 0},   
                {2, 0},
                {3, 0}
              }
We are using following profile Cfg:
profileCfg 0 60 15.65 7 15.6 0 0 256 1 60 6000 0 0 48
We are observing a crash at AoAProcHWA_peakGrouping() funtion in file aoa2dprochwa.c at line no. 346.
This crash is random in behavior and happening at different time instances for multiple runs of same code.
PFA the screenshot in the CCS debug window
Please let me know anything that I am doing wrong or what could be done to solve this issue.

Thanks & Regards,
Prathith

  • 0
    TI__Mastermind 23725 points

    Hello

    Have you be able to see how far  inside the function the execution happens.

    Considering you have introduced changes to the code you may want to see how the changes are  affecting the AoA algorithm.

    Thank you,

    Vaibhav

  • 0 in reply to Vaibhav Mahimkar
    Prodigy 70 points

    It is stopping at continue statement

    if (detList[i].peak < threshold)
            {
                continue;
            }
    the debugged value in ccs is as follows
    At first debugging process I got to know that the index accessing the detList is going out of bounds. Is this happening due to corruption of memory by any chance?
    I feel that the input detList buffer given to this function is shared among other processes. Please let me know how to debug further
    Also, to debug among the changes we have made, this crash is happening randomly that I cannot conclude anything.
    Thanks & Regards,
    Prathith
  • 0 in reply to V Prathith Vasudev
    TI__Genius 14415 points

    Hello,

    Can you please confirm whether you are using one of the EVM antennas or a custom antenna design? I am gathering a little more information about these functions and will get back to you.

    Regards,

    Jackson

  • 0 in reply to Jackson Thomas
    Prodigy 70 points

    I am using a custom antenna design.

    regards,

    Prathith

  • 0 in reply to V Prathith Vasudev
    Prodigy 70 points

    Though i checked the same thing on 6843 ODS-ISK board. the same behavior is happening.

  • 0 in reply to V Prathith Vasudev
    TI__Genius 14415 points

    Hello,

    How have you updated your antenna design? It seems that maybe you are overrunning the index in the variable below? Or one of these indexes or maxPoints is being set too high and allows to use too much memory.

    if (detList[i].peak < threshold)

    Can you keep some list of numDetectedObjects to make sure it makes sense and isn't set improperly? I would also check the maxNumExpectedPeaks and its calculation from the antenna params to verify that you are not somehow detecting more than this.

    Regards,

    Jackson

  • 0 in reply to Jackson Thomas
    Prodigy 70 points

    Hi Jackson,

    I have updated my antenna design in antenna_geometry.c file. I put a check on numDetectedObjects, the maximum value it is having is 256. and also maxNumExpectedPeaks  is set to 256. and I assume the variable is calculated in this manner 

    aoaHwaObj->maxNumExpectedPeaks = DPU_AOAPROCHWA_NUM_ANGLE_BINS * aoaHwaObj->numAntCol;
    I see nothing that can take the iterations to so much large value. suddenly the i and j values are becoming very high.
    Regards,
    Prathith
  • 0 in reply to V Prathith Vasudev
    TI__Genius 14415 points

    Hello Prathith,

    Let me look into this a little more and get back to you by next week. I am not surprised that the iterations become very high if you try to log each frame.

    Have you been able to isolate anything that is correlated with this error? You say it is random but can you do anything to reproduce it more often? Extra clutter or more dynamic points or something?

    Also, you say you checked the same thing happens with the ODS board. Did you use the same antenna configuration in SW with this board? Does the issue go away if using the default antenna configuration?

    Regards,

    Jackson