• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 29 subscribers
  • Views 301 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.

AWR1642BOOST-ODS: Occupancy detection use case

Shihyu Hung
Intellectual 260 points
Part Number: AWR1642BOOST-ODS
Other Parts Discussed in Thread: TEST2, AWR6843AOP

Hi, TI team,

I ran the obstacle detection project (using gui of automotive toolbox 3.4 lab-4) with AWR1642-ODS, but I found that the detected bounding boxes show really unstable.

And here is the description of my experimental setup, also the video recorded while testing:

2 people sat in front of the EVM kit at about 1 meter away, and about 60 degrees within each other. (almost static)

Do you have any recommendation to improve the stability?

(I also retry again using two corner reflectors instead, the signal seems to be more strong and stable, but this is not feasible for real case.)

Many thanks,

Shihyu.

  • 0
    TI__Guru*** 149044 points

    Hi

    The boxes are used to show clusters not detected points.

    I think what you should be interested in are the detected points not the clusters. So you may want to disable the Clusters in the GUI settings

    Thank you
    Cesar

  • 0 in reply to Cesar
    Intellectual 260 points

    The clusters are generated according to the points, and apparently those points are very unstable. No matter the positions in same frame or in sequence.

    Shihyu.

  • 0 in reply to Shihyu Hung
    TI__Guru*** 149044 points

    Hi,

    Can you please try the following use case:

    1 person static at 1 m, at 30 degree from bore sight

    1 person static at 1.5 m at -30deg from bore sight

    The detection should be robust

    Thank you

    Cesar

  • 0 in reply to Cesar
    Intellectual 260 points

    Hi, Cesar,

    It looks much more stable to detect 2 people after using the setup from your recommendation. Thanks a lot.

    And I wanna ask that why the second setup is much more stable than previous one? (which people were lined up in same row)

    Is this scenario not suitable for the AWR1642-ODS? Since according to the spec. of Azimuth resolution, this kit  indeed support this scenario, and I can't not figure out why my test went wrong?

    And may I ask the difference and the ability of AWR1642-ODS and AWR6843AoP? One for Occupancy detection and another one is for In-cabin presents detection, but as I know, these two tasks are quite similar. What do you think of these two kits?

    Thanks for your help,

    Shihyu.

  • 0 in reply to Shihyu Hung
    TI__Guru*** 149044 points

    The reason why the setup is more stable is that we are detecting static objects in different range bins.

    It is important to understand that when objects are static, detection is only based on range and angle information since the velocity is zero.

    So, when the range information is the same, the only the angle information can be used to detect. So, the detection is less robust.

    If the objects are moving, doppler information will be used for detection and detection will be more robust.

    Shihyu Hung said:
    And may I ask the difference and the ability of AWR1642-ODS and AWR6843AoP? One for Occupancy detection and another one is for In-cabin presents detection, but as I know, these two tasks are quite similar. What do you think of these two kits?

    The most important difference between these two sensors is that the 68xx works with 60GHz and 16xx works with 77GHz.

    The bandwidths available for different applications may depend on countries. You would need to check with regulatory organizations in your own country.

    The second most important difference is that 68xx has integrated antennae. So this can't be changed

    The third most important difference is that 68xx has 3 Tx. The third Tx is used to provided elevation information

    thank you

    Cesar