• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 1196 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.

Original question:

IWR1843BOOST: implementation of ValidateChecksum matlab function in python for 2D People Counting Lab

IWR1843BOOST: People Counting Demo TLV mapping.

Angelo Claudio Re
Intellectual 300 points
Part Number: IWR1843BOOST
Other Parts Discussed in Thread: IWR1843, IWR6843, , IWR6843ISK, MMWAVEICBOOST

Hi,

I have a doubt for what concerns the mapping among different TLV packets that follow the Frame Header for the 2D People Counting Lab.

I'm using SDK 03.04.00.03 and the mmWave Industrial Toolbox 4.0.0.

There are 3 types of TLV packets:

  • POINT CLOUD TLV (6)
  • TARGET LIST TLV (7)
  • TARGET INDEX TLV (8)

Is the mapping among them strictly unique ? I mean while I'm processing the data coming from the UART port of the device I'm going to receive as a sequence, after the Frame Header, a POINT CLOUD TLV, a TARGET LIST TLV and a TARGET INDEX TLV that are related to the same point cloud detected. Is it right ?

If not so, in the case I can receive as output of the UART port multiple consecutive POINT CLOUD TLV, or TARGET LIST TLV, or TARGET INDEX TLV (e.g. 20 POINT CLOUD TLV, 2 TARGET LIST TLV, 3 TARGET INDEX TLV) how could it be possibile to understand the way in which they are related to each other ?

Kind regards

Angelo Claudio Re

  • 0
    TI__Mastermind 23725 points

    Hello

    Please mention the exact lab/Demo which you are trying to interpret.  As you may notice in the UG for each example the set of TLVs are repeated once every frame.

    You want to check the  uart out section of the  source code that runs on device to understand how  data is constructed before sending out.

    Hope this helps.

    Thank you,
    Vaibhav

  • 0 in reply to Vaibhav Mahimkar
    Intellectual 300 points

    Hi,

    I'm using SDK 03.04.00.03 and the mmWave Industrial Toolbox 4.0.0.

    The lab is this one : https://dev.ti.com/tirex/content/mmwave_industrial_toolbox_4_0_0/labs/people_counting/68xx_people_counting/docs/pplcount_user_guide.html

    I read another time the documentation regarding the Output Data format, and I got the following considerations about the 3 types of TLV packets:

    • POINT CLOUD TLV (6) - It is used to represent a point detected in the scope of the device.
    • TARGET LIST TLV (7) - It is used to represent an object detected by the device, and not a single point. This object will be tracked in the space as long it'll remain in the range of the device.
    • TARGET INDEX TLV (8) - It is used to specify which points of the POINT CLOUD TLV are associated to the object detected through the TARGET LIST TLV.

    Is the above explanation of the 3 TLV packets correct ? Am I interpreting them in the right way ?

    Also, the main difference between the 2D People Counting Lab and the out of the box demo is that the former tracks objects in the space instead the latter only detects points in the space and if you want to do tracking or object detection you have to develop them on your PC after collecting the data of the device, or modify the C source code as you please. Is it right ?

    Finally, the UART  OUT SECTION for the lab that I linked is the one implemented in the following source code file ? ($HOME/mmwave_industrial_toolbox_4_0_0/labs/people_counting/68xx_people_counting/src/mss/task_mbox.c)

    Kind regards,

    Angelo Claudio Re

  • +1 in reply to Angelo Claudio Re
    TI__Genius 14415 points

    Hello Angelo,

    Your description of the 3 TLVs is correct. You are also correct about the OOB demo not including the tracker output. This can be added into the OOB demo (or done on an external host/PC if desired), as the tracking code is all included in the SDK. An example of this would be in the Area Scanner lab, where it includes the tracker on top of an OOB demo processing chain. You can see how the gtrack library is included in that demo.

    The UART transmission is handled by the UART driver from the SDK.

    Also, is there a reason you are using the IWR1843 for people counting? Is it specifically for an automotive use case? Otherwise I would suggest moving to the IWR6843 which will support the updated 3D people counting demos.

    Regards,

    Jackson

  • 0 in reply to Jackson Thomas
    Intellectual 300 points

    Hi Jackson, thanks for your reply.

    I'm using the IWR1843 because in my Master Thesis I'd like to create a system able to classify, detect and track drones in a given space and finally I'm going to perform a benchmark among 3 devices, the IWR1843BOOST, the IWR1842BOOST and the IWR6843ISK + MMWAVEICBOOST.

    Regards,

    Angelo Claudio Re

  • 0 in reply to Angelo Claudio Re
    TI__Genius 14415 points

    Ok I understand. The performance may differ based on the software offering currently available. The 1843 and the 6843 should have largely the same performance, as they have the same TX and RX antennas and same processing, just slightly different frequency. This is usually determined by regulations on which frequency band can be used. But it could be very useful to have the specific performance differences between the two devices. 

    Regards,

    Jackson