• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 31 subscribers
  • Views 351 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.

IWR6843ISK: Long range people detection can't reach performance expectation

Fabio Takeuchi
Prodigy 10 points
Part Number: IWR6843ISK

We are trying to use the development kit for long range people detection.

We followed the guide using firmware C:\ti\<mmwave_industrial_toolbox_install_dir>\labs\long_range_people_detection\68xx_long_range_people_det\prebuilt_binaries\long_range_people_det_68xx_demo.bin

Chip configuration C:\ti\<mmwave_industrial_toolbox_install_dir>\labs\long_range_people_detection\68xx_long_range_people_det\chirp_configs\people_detection_and_tracking_100m_2D_advanced.cfg

And GUI C:\ti\<mmwave_industrial_toolbox_install_dir>\labs\traffic_monitoring\18xx_68xx_traffic_monitoring\gui\tm_visualizer.exe

As a result, we've got the longest distance as shown in the image below.

We are too far from performance expectation shown in the table below

We've got something near from 20 meters and the expected performance says more than 50 meters.

What should we need to do to get the expected distance?

  • 0
    TI__Mastermind 23725 points

    Hello,

    For Debug in general:

    Would it be possible for you to try these measurements in an open field so that one can correlate between the scene and objects.

    A good test would be to start from close to the sensor and walk away so that you can check results side by side.

    A few questions:

    Fabio Takeuchi said:
    We've got something near from 20 meters and the expected performance says more than 50 meters.

    Are there no objects/people in the 50m area where the demo detects objects/people?

    What does the scene look like for the missing and false detected objects ?

    Thank you,

    Vaibhav

  • 0 in reply to Vaibhav Mahimkar
    Prodigy 10 points

    Thank you Vaibhav for your answer. I should have presented more details.

    This is the scenery where we've got the signal. Unfortunately I was right behind the grid as shown in image below.

    The radar was above and after the grid, but I haven't taken a picture.

    In the image showed in the previous post I was the point marked by the arrow

    I guess the others point clouds are ghost signals or trees.

    The video below shows what happened during the test. The circle shows where I was and the arrow points where I was by radar perspective. As you can see, the target stops a little more than 20 meters.

    About your questions.

    There could be trees in 50 meters area, but not others objects nor people.

    Thank you again Vaibhav.

  • 0 in reply to Fabio Takeuchi
    TI__Genius 14415 points

    Hello,

    It is not clear from the video where you actually are after reaching 20 meters, you become obscured in the camera by the grid.

    Is there a reason you are using the 2D configuration? The tracking should be a little better in the 3D configuration. I would suggest to try the 3D 50m configuration to see if that helps. The 100m profile has a much lower frame rate.

    Regards,

    Jackson

  • 0 in reply to Jackson Thomas
    Prodigy 10 points

    Hi Jackson,

    Actually there is no reason to use that configuration. I just thought that would be the configuration for long range detection.

    So I'll try again changing this,

    Thank you.