IWR1443BOOST: Frame Num and x, y, z Coordinates

Part Number: IWR1443BOOST
Other Parts Discussed in Thread: IWR1443

Tool/software:

Hello E2E Experts,

Good day.

I tried executing the code on the recorded .dat file from demo visualizer. It keeps printing in a loop about how many detected objects are there. My requirement however, is to get the frame num and x,y,x coordinate of each detected object.

How can I do this?

Regards,

TICSC

  • Hello, 

    I'm not sure exactly what code you are referring to. Can you please provide a bit more detail? 

    Also, if you're running the out of box demo and simply looking to capture the detected objects per frame and their xyz coordinates then you could also try using our python based visualizer in the Radar Toolbox. It has the option to save the output data to a file as well as replay the data. Also the visualizer source code is open for you to view/modify. Here is a link to the User Guide for that visualizer: mmWave Applications Visualizer

    Regards,

    Josh

  • Hello Josh,

    Good day.

    I have the device IWR1443 boost, and I used the online demo visualizer version 2.1.0. When I click the Start record button and record the data, I get a .dat file. How can get the frames of point-cloud out of it? Please help me. I python script would be very helpful that takes the .dat file and gives me the numpy array of frames of point cloud. 

    As mentioned,  my requirement is to get the frame num and x,y,x coordinate of each detected object.

    Regards,

    TICSC

  • Hello, 

    Okay understood. Thanks for the additional information. There is another e2e thread on the same topic: https://e2e.ti.com/support/sensors-group/sensors/f/sensors-forum/985107/awr1642boost-mmwave-demo-visualiser-captured-data 

    Best Regards,

    josh

  • Hello Josh,

    Good day.

    I tried executing the script but I am constantly getting this:
    "
    Saving message 899 as raw string
    CorruptMessage: TLV type not recognized
    Saving message 900 as raw string
    CorruptMessage: TLV type not recognized
    Saving message 901 as raw string
    "

    I used the script with recorded .dat file from this link: "https://dev.ti.com/gallery/view/mmwave/mmWave_Demo_Visualizer/ver/2.1.0/
    I have also attached the .dat file that I obtained from the above link. The SDK version of my device is 2.1. Also, I show the sample .json in the zip file in the thread and I want the data like that, where I can get the x,y,z coordinates with frames.
    Regards,
    TICSC
  • Hi TICSC, 

    I tried executing the script with the data you shared and I do not get any errors. You could check if your python version may be causing the issue somehow? I'm using version 3.7.4. Here is the parsed JSON file: 
    /cfs-file/__key/communityserver-discussions-components-files/1023/parsed.json

    Regards,

    Josh

  • Hello Josh,

    Good day.

    I tried python 3.7.4 and got the same error. Can you please have an online meeting with me so that you can see what the issue on my side is. Also, can you please send me the exact python env you used to parse the *.dat file sent by me in the last email ? You can also suggest me any other device of TI that can directly give me the X,Y,Z coordinates of all detected objects per frame to make my work easy. 

    Regards,

    TICSC

  • Hello, 

    Sorry for the delayed response. That would be outside the scope of these forums. As the linked thread mentioned we can't provide any debugging support for this script. 

    "Please note that this script is not officially released, and may never be released. As this is an experimental script, I will not be able to guarantee that it works for every use-case, and our team won't be able to support this on E2E."

    You can also suggest me any other device of TI that can directly give me the X,Y,Z coordinates of all detected objects per frame to make my work easy. 

    The mmWave Applications Visualizer supports saving the parsed output data to a JSON file (including x,y,z data for the point cloud) however it does not support 1443 since that is an older device. You could switch to one of the xWR6843/xWR1843 EVMs or xWRL6432/xWRL1432 EVMs which are all supported by that visualizer. Or you could try to modify the visualizer source code to support parsing the TLVs output by 1443. 

    Best regards,

    Josh