• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 594 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.

IWR6843AOPEVM: Ghost points in moving robot

Felipe Xavier
Prodigy 30 points
Part Number: IWR6843AOPEVM

Greetings,

I have recently being testing with the IWR6843AOPEVM for obstacle avoidance in a mobile robot. The sensor is mounted on top of the robot (2m up) in standalone version and although it works well in general, there seems to be an frequent detection of "ghost" points (the points appear at varied distances and angles). I read that the bouncing of the signal in corners can cause such false positives but the test area is quite open, 6m x 6m and I imagine this could minimize the issue. I have tried decreasing the sensor sensitivity but that affects the obstacle detection capabilities. Furthermore, I have also implemented some KNN algorithms to detect noise but due to the sparse cloud that also affects the obstacle detection capabilities.

Do you know what could be the cause of the false positives? Is such a behavior expected? What would you recommend doing in such a scenario?

Thanks in advance.

Kind regards,

Felipe Xavier

  • 0
    TI__Mastermind 20755 points

    Hi Felipe,

    Is it possible for you to share the chirp config you are using.

    Thanks

    Yogesh

  • 0 in reply to Yogesh Siraswar
    Prodigy 30 points

    Hi Yogesh,

    Of course, I forgot to add it before and also forgot to add that the desired obstacle detection range is between 1 and 4m from the sensor.

    sensorStop
    flushCfg
    dfeDataOutputMode 1
    channelCfg 15 7 0
    adcCfg 2 1
    adcbufCfg -1 0 1 1 1
    profileCfg 0 60 43 7 40 0 0 100 1 224 7000 0 0 30
    chirpCfg 0 0 0 0 0 0 0 1
    chirpCfg 1 1 0 0 0 0 0 2
    chirpCfg 2 2 0 0 0 0 0 4
    frameCfg 0 2 16 0 33.333 1 0
    lowPower 0 0
    guiMonitor -1 1 0 0 0 0 0
    cfarCfg -1 0 2 8 4 3 0 8.0 0
    cfarCfg -1 1 0 4 2 3 1 8.0 1
    multiObjBeamForming -1 1 0.3
    clutterRemoval -1 0
    calibDcRangeSig -1 0 -5 8 256
    extendedMaxVelocity -1 0
    lvdsStreamCfg -1 0 0 0
    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 4 99 0
    CQSigImgMonitor 0 111 4
    analogMonitor 0 0
    aoaFovCfg -1 -60 60 -30 30
    cfarFovCfg -1 0 0.5 5
    cfarFovCfg -1 1 -8.02 8.02
    sensorStart

    Thanks for the quick reply.

    Kind regards,

    Felipe Xavier

  • 0 in reply to Felipe Xavier
    TI__Mastermind 20755 points

    Hi Felipe,

    There could be multiple reason for ghost points, if there is multi-path reflection in the scene, it could be ceiling, floor, and double reflection and etc.  

    Here are couple of points to debug:

    1)  Given a CFAR threshold, there will be always some false detection.  You can increase the threshold or use clustering/tracking type of up level method to filtering those false detection.   

    2) You can also increase the number of loop to increase the system SNR and then also increase the threshold so that it will not affect the obstacle detection capabilities. 

    3) Check if there is any glitch in the noise, it can be checked by analyzing the data with an empty scene or aiming the sensor outside to the open air.

    Thanks

    Yogesh

  • 0 in reply to Yogesh Siraswar
    Prodigy 30 points

    Hi Yogesh,

    I had previously tested in an empty scene and there seemed to be no glitch in the noise. I will try the other given suggestions.

    Thank you for the recommendations.

    Kind regards,

    Felipe