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.

IWR1642: Error covariance matrix and Gating Gain understanding

Part Number: IWR1642

Hello,

I'm working with IWR1642.

Could you please explain for which purposes there is Error Covariance Matrix and Gating function gain in output packet?

Could these parameters be used to object classification?

  • Hi, 

    Where did you see the information about Error Covariance Matrix and Gating function gain?  I do not think they are included in the SDK out-of-box demo.  Those parameter can be related to tracking module.  Tell me let me know which demo you are trying to use that has these information?

    Best,

    Zigang

  • Hi Zigang,

    We use people counter demo. We almost didn't modified original firmware form SDK (we only disabled cloud of points). 

    For now IWR1642 sends Target list to uart and we are working with these data. 

    Here is description from People Count User Guide:

    Question is how to use these data.

    Best regards,

    Leonid

  • Hi, Leonid:

    These information is generated by tracker and used in the visualizer.   

    A bottom panel with a Gating and Association plot.

    • This plot visualizes the result of the tracking algorithm.
    • The colored points represent the detection points (black points in previous frame's Point Cloud plot) which are associated to a specific track. Unassociated points that do not belong to a track are not plotted.
    • The colored circular ring is centered over the centroid of of the tracked object. The diameter of the ring is related to the variance in location of the tracked object's detection points.

    You may find more information in C:\ti\<mmwave_industrial_toolbox_install_dir>\labs\people_counting\68xx_people_counting\docs\pplcount_customization_guide.pdf

    Best,

    Zigang

  • Hi Zigang,

    thank you for explanation.

    Our device which is connected to IWR1642 via UART use the same data packet.

    We capture X,Y and other data and use it for our own application and visualization.

    I want to know is it possible to use EC and G fields for getting some information about object size to make object classification (human, animal, car etc) ?

    Or how can we use this information?

    Best regargs,

    Leonid

  • Hi Leonid,

    EC and G are not an accurate way to determine size of the object. The tracker returns as index TLV, which allows you to tell which points are associated with the tracker. I recommend analyzing the points associated with a target to determine the size of the point cloud, which will give a rough estimation of object size.

    Regards,

    Justin

  • Hi Justin,

    Thank you for your answer. We will try analize the point of cloud.

    Best regards,

    Leonid