• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 32 subscribers
  • Views 218 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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: Calibration and Object Detection Issue with IWR6843AOP-EVM

Kai
Prodigy 100 points
Part Number: IWR6843AOPEVM

Hello,

I'm reaching out regarding an issue I've encountered while using the IWR6843AOP-EVM for extracting the compRangeBiasAndRxChanPhase parameters, as outlined in the "Detection Layer Parameter Tuning Guide for the 3D People Tracking Demo – Rev 3.1."

To provide context, I've followed the prescribed steps meticulously. This includes utilizing the demo binary and calibration profile from "mmwave_sdk_03_05_00_04\packages\ti\demo\xwr68xx\mmw" and adhering to the calibration procedure documented in "mmwave_sdk_03_05_00_04/packages/ti/demo/xwr68xx/mmw/docs/doxygen/html/index.html." Additionally, I've placed a metal reflector measuring 12cm by 15cm, positioned at 1m from the sensor during calibration.

Enclosed with this email, you'll find visual representations in the form of Range-Profile and Azimuth-Range-Heatmap graphs. Furthermore, I've included a snippet of the calibration parameters printout for your reference.

compRangeBiasAndRxChanPhase 0.0731140 0.23532 -0.81952 -0.37357 0.75186 0.39874 -0.62717 -0.28342 0.72424 0.23337 -0.95166 -0.44833 0.89386 0.41525 -0.76297 -0.31064 0.84314 0.21487 -0.89243 -0.39841 0.82614 0.39572 -0.70679 -0.29651 0.78278

To refine the configuration, I've averaged parameters over 10 samples and integrated them into the original config file of the People Counting Project located in "mmwave_industrial_toolbox_4_12_0\labs\People_Counting\3D_People_Counting\chirp_configs\ AOP_6m_default.cfg." Additionally, I've adjusted the sensor position within the config file to "sensorPosition 3.0 45 35” to accommodate a 35-degree tilt of the EVM module.

However, despite these measures, I've encountered an unexpected discrepancy in object detection. Specifically, when evaluating the height of a "13cm X 13cm" PC fan positioned vertically on the floor at coordinates X=1m and Y=5.4m, the detected height registers as Z=43cm. This contrasts starkly with the anticipated maximum height of 12cm.

Any insights or guidance you can offer to solve this issue would be appreciated.

Regards,

Kai

  • 0
    TI__Expert 8790 points

    Hi 

    Thanks your for your query. Please allow us a couple of days to respond

    Regards

  • 0 in reply to Abhishek Karkisaval
    TI__Mastermind 32310 points

    Hi Kai,

    Do you get different height results when you don't run the calibration? 

    Best,

    Nate

  • 0 in reply to Nathan Block
    Prodigy 100 points

    Hi Nate,

    Without the calibration, the reported height is 60cm.

    Cheers,

    Kai

  • 0 in reply to Kai
    TI__Mastermind 32310 points

    Hi Kai,

    Is the sensor placed 3 meters in the air? That's what your CLI command would indicate:

    Kai said:
    sensorPosition 3.0 45 35

    Best,

    Nate

  • 0 in reply to Nathan Block
    Prodigy 100 points

    Yes, the sensor is placed at 3 meters in the air with a 45-degree azimuth rotation and a 35-degree tilt, to make a corner mount configuration.

    Regards,

    Kai

  • 0 in reply to Kai
    TI__Mastermind 32310 points

    Thank you Kai. 

    Is the range estimation accurate? What is your range resolution? An error of 30 cm may not be unreasonable if it's mostly in the angle domain, especially for a moving fan. You could also try turning clutter removal off, and trying to localize a corner reflector instead. This should eliminate any of the potential increased error from the fan's velocities.

    Is the entire fan 13cm x 13 cm in size?

    Best,

    Nate

  • +1 in reply to Nathan Block
    Prodigy 100 points

    Hi Nate,

    After some troubleshooting, I relocated the sensor to a different room, maintaining a consistent height of 3 meters. I'm pleased to inform you that this action effectively resolved the issue we were experiencing.

    With the sensor now in its new location, we are consistently obtaining accurate height readings in every corner of the room, with a deviation of only approximately ±10cm. 

    Upon further inspection, it appears that the presence of exposed metal cable ducts on the ceiling was indeed the root cause of the issue. By addressing this potential interference, we've successfully mitigated any disruptions to the sensor's functionality.

    Thank you for your patience and support throughout this process.

    Regards,

    Kai