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.

IWR6843ISK: IWR6843ISK : Area scanner Lab and 3D people counting Lab for detection of Objects and Humans: Looking for Deep understanding on Radar Tracking

Part Number: IWR6843ISK

Hello Experts, 

I am using IWR6843ISK for my evaluation on Area scanner Lab and 3D people counting Lab for detection of Objects and Humans and compare the results. It would be great if I could get more understanding on the Labs. That would be much helpful. 

Where I can find more information about the Area Scanner Lab: 

  • Static Point Cloud :
    • How is the static detection processing done?
    • Will I be able to detect the presence of Humans with this information from the Area Scanner? 
  • Dynamic Point Cloud: 
    • Where I can get more and deep understanding on traditional range-doppler processing chain?
    • Where can I find logic and algorithm or clear steps on static clutter removal enabled?
  • Tracked Objects: 
    • Where can I find the clear information on how does GTRACK algorithm work ?
    • What is the block diagram of GTRACK algorithm? 
    • What type of grouping is done? Which type of Clustering is done here for Tracked Objects? 
    • It is mentioned that "The dynamic point cloud is fed as input to the tracking algorithm, which groups and tracks points across the scene". I have extracted the dynamic points data object list from TLV 1 and I have run for longer time for more than 30 secs. I haven't got any Tracked Data even after 30–40 secs after an object is moved dynamically in the scene and I haven't got any info from TLV 10, what could be the reason for no tracked data even though there are Dynamic Objects? 
    • What brings a Tracked object data point ? What is the dependent factor between Tracked Objects and Dynamic Objects ? 
    • How they are grouped ?
    • How much time is needed for a Dynamic object to be get tracked?  Could you please give more information on this ?

Where I can find more information about the PEOPLE COUNTING LAB: 

  • What TLV numbers are assigned for Static objects, Dynamic Objects, Tracked Object?
  • Currently, I am using .cfg files in my standalone data parsing (Data acquisition) application which will parse the data coming to the ISK sensor, I have given the .cfg file from Area scanner and extracted Static objects, Dynamic Objects, Tracked Objects. If I could know the TLV numbers and API for the People counting, I wanted to extract data from the 3D people Counting Lab ? 

Looking forward for your response; 

Wishing you a nice day ! 

Best regards,

Shravya 

  • Hi Shravya,

    • Where can I find the clear information on how does GTRACK algorithm work ?
    • What is the block diagram of GTRACK algorithm? 
    • What type of grouping is done? Which type of Clustering is done here for Tracked Objects? 
    • It is mentioned that "The dynamic point cloud is fed as input to the tracking algorithm, which groups and tracks points across the scene". I have extracted the dynamic points data object list from TLV 1 and I have run for longer time for more than 30 secs. I haven't got any Tracked Data even after 30–40 secs after an object is moved dynamically in the scene and I haven't got any info from TLV 10, what could be the reason for no tracked data even though there are Dynamic Objects? 
    • What brings a Tracked object data point ? What is the dependent factor between Tracked Objects and Dynamic Objects ? 
    • How they are grouped ?
    • How much time is needed for a Dynamic object to be get tracked?  Could you please give more information on this ?

    See the PDF here in the People Counting docs. The tracker for people counting and area scanner is the same.

    https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.00.00.26&node=A__APAFl-amXsEcKJ4q3oOjFw__radar_toolbox__1AslXXD__1.00.00.26

    • Static Point Cloud :
      • How is the static detection processing done?
      • Will I be able to detect the presence of Humans with this information from the Area Scanner? 
    • Dynamic Point Cloud: 
      • Where I can get more and deep understanding on traditional range-doppler processing chain?
      • Where can I find logic and algorithm or clear steps on static clutter removal enabled?

    The Area Scanner dynamic point cloud is based off our Out-of-box example at the detection layer. See the SDK documentation at C:\ti\mmwave_sdk_03_06_00_00-LTS\docs.

    You can learn more about static clutter removal here. https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.00.00.26&node=A__AO7sunER70ma-iu3FCioXQ__radar_toolbox__1AslXXD__1.00.00.26

    • What TLV numbers are assigned for Static objects, Dynamic Objects, Tracked Object?
    Currently, I am using .cfg files in my standalone data parsing (Data acquisition) application which will parse the data coming to the ISK sensor, I have given the .cfg file from Area scanner and extracted Static objects, Dynamic Objects, Tracked Objects. If I could know the TLV numbers and API for the People counting, I wanted to extract data from the 3D people Counting Lab ? 

    See the UART Output Data Format section of https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.00.00.26&node=A__AI86Gas91BD7pq5ikLOshQ__radar_toolbox__1AslXXD__1.00.00.26

    Best,

    Nate

  • Hello Nathan, 

    Could you give me the key differences between OBB demo with Area Scanner and 3D people counting in terms of:

    • What kind of Clustering algorithm is used for Tracking the Objects and Humans from both demos:, I have read the document which says G Track algorithm, but there are not much info on it. Could you mention the more information on Clustering process. 
    • Detection Algorithm: What is the Detection algorithm used ?

    I have faced issues while using Area Scanner Lab to detect the Objects, A lot of times I was getting dense dynamic points, but there are no tracked objects when I parsed the data. Sometimes, I'm getting both Dynamic data and tracked data. 

    More explanation about tracking objects from Area scanner would help as there are issues while testing the Demo. 


    Could you please address this [problem|? Let me know if you need any further info from my end.

    Best Regards
    Shravya 

  • What kind of Clustering algorithm is used for Tracking the Objects and Humans from both demos:, I have read the document which says G Track algorithm, but there are not much info on it. Could you mention the more information on Clustering process.

    See our documentation on gtrack here.

    https://dev.ti.com/tirex/explore/node?node=A__APAFl-amXsEcKJ4q3oOjFw__radar_toolbox__1AslXXD__LATEST&r=1AslXXD__1.00.00.26

    • Detection Algorithm: What is the Detection algorithm used ?

    See our documentation on the out-of-box toolchain at the following file location.

    C:/ti/mmwave_sdk_03_06_00_00-LTS/packages/ti/demo/xwr68xx/mmw/docs/doxygen/html/index.html

    I have faced issues while using Area Scanner Lab to detect the Objects, A lot of times I was getting dense dynamic points, but there are no tracked objects when I parsed the data. Sometimes, I'm getting both Dynamic data and tracked data. 

    More explanation about tracking objects from Area scanner would help as there are issues while testing the Demo. 


    Could you please address this [problem|? Let me know if you need any further info from my end.

    See our tracker layer tuning guide here.

    https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.00.00.26&node=A__AAwhSXiV2o4yQlaiTsMgQg__radar_toolbox__1AslXXD__1.00.00.26

    Best,

    Nate