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.

AWR1443BOOST: Best device configuration for human posture representation

Part Number: AWR1443BOOST
Other Parts Discussed in Thread: AWR1443, IWR1443, IWR6843

Hello TI community,

I'm writing this post to ask for a piece of advice about the best configuration (config file) to set onto the TI AWR1443 for best human posture representation, in my case i'm using the OOB image on the card, and i'm getting some representation of moving objects in a point cloud data (.pcd) representation using the ROS driver developed by the TI team.

I added some modifications on the driver to get the point cloud at a single timestamp (one frame), but i'm getting a very small number of points in the same frame (about 4-8).

So my questions are : 

  • Is IWR1443 the best device to use in this kind of applications?
  • Is the image i'm flashing it with the best one to use in this application?
  • Is there any config file you consider you think is better then what i'm using (i'm using the zone occupancy config : industrial toolbox 3.6.2)?

Thank you in advance for your help.

  • Hi, 

    Can you please let me know the target application? Is it intended to track moving person or some one seated? 

    Regards, 

    Sudharshan K N 

  • Hi Sudharshan,

    Thank you for you answer,

    The target application is to track the person's posture in movement. In other words, the final aim is to detect the dangerous and sudden movements (like fall). If i can at the same time track the moving and static presence this will be better. But the most interesting for me is to track moving bodies.

    Abdellah.

  • Hi, 

    Thanks for the clarification! We do have few demos for the fall detection using our 60GHz sensors. Please refer to the experiment https://dev.ti.com/tirex/explore/node?node=AI.qfRfVzNcQ5es7TsPfuw__VLyFKFf__LATEST for further details. 

    Regards, 

    Sudharshan K N 

  • Hi Sudarshan,

    Thank you for your help,

    I've already seen that, but for know i'm using the IWR1443 77-81 Ghz, and that's why i asked the above questions.

    Appreciate your help.

    Abdellah.

  • Hi Abdellah,

    Our fall detection experiment that Sudharshan references above is based on our 3D people counting implementation. This lab is based on our IWR6843 device which uses 60GHz. 

    However, that does not mean fall detection is not possible using the 77-81GHz band. In older releases of our toolbox you will find IWR16xx based people counting algorithms or you could modify the existing IWR6843 lab to work on IWR1443. 

    The easier way to get started would be to switch to a device which already has source code working on it like the IWR6843. Starting with the experiment linked above will greatly decrease your development time.

    Thank you,

    Angie

  • Hi Angie,

    I insisted on the use of iwr1443 because i feel that i'm really close to the goal, i've already done body representation but with some noise (unexciting points : something like fixed points near the radar). In fact i need just some orientations on the use of the config file to get best representation (maximum range and without noise and with maximum point cloud representation of the body).

    Thank you for you explanation and advices, i'd ask just one more question. what's the difference between AOP, ODS and ISK models for the IWR6843 EVM, what is the role of the carrier board, And what is the best device for my use case application.

    Have a good day.

    Abdellah.

  • Hi Abdellah,

    Thank you for explaining! You can take a look at our source code and find ways to apply the logic onto IWR1443.

    The difference between these modules is in antenna design. The AOP has an antenna on package, the ISK is ideal for wall mount use cases, and the ODS is ideal for overhead use cases. The full antenna details can be found here

    Thanks,

    Angie