• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 31 subscribers
  • Views 227 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

AWR1843BOOST: raw data processing

Wembanyama
Intellectual 600 points
Part Number: AWR1843BOOST
Other Parts Discussed in Thread: DCA1000EVM

hi,

I want to know, whether I can process the raw data captured by dca1000 board using machine learning method, to fit objects' trace?

whether TI support some demo? 

Regards,

Wembanyama

  • 0
    TI__Expert 7755 points

    Hey Wembanyama, 

    Thanks for reaching out regarding your query into applying ML to raw data. First of all, could you describe in greater detail what fitting an object's trace is? Are you suggesting being able to differentiate objects and track them? If so, TI does have a demo, the Medium Range Radar demo in the Radar Toolbox, for estimating and tracking the distance and velocity of objects. Note that this demo does not use ML.

    We also have a gesture recognition demo that uses ML to classify different detected movements in front of a radar devices into gestures. However, this is for xWR6843AOP or xWR6843ISK-ODS. I think its better to look at this as a reference for how ML can be applied on one of our devices to help you develop your own demo for AWR1843BOOST. The ML model used here does not use raw data and is trained on various extracted features from the radar device - more details can be found in the user guide.

    If you could clarify more details on your application, then we might be able to point you in a better direction. Let me know if you have any other questions or if these demos helped.

    Regards,

    Kristien

  • 0 in reply to Kristien Everett
    Intellectual 600 points

    hi,

    Very appreciate for your response. my question is:

    I used 1843+DCA1000evm to detect one object moving in a certain range of motion, set 1Tx4Rx, and gathered the raw BIN data from dca1000evm.  finally I correctly converted the BIN data to corresponding I/Q decimal data， for example, 65+121i, -3-27i,......

    Now, I want to use these I/Q decimal data applying to some ML model or ML algorithms, so that I can gain my object's range and azimuth. 

    ------------------------------------------------------------------------------------------------------------------------------

    Kristien Everett said:
    TI does have a demo, the Medium Range Radar demo in the Radar Toolbox, for estimating and tracking the distance and velocity of objects
    Kristien Everett said:
    We also have a gesture recognition demo that uses ML to classify different detected movements in front of a radar devices into gestures

    unfortunately, the two website link is 403 Forbidden, how can I get into the two website link.

    Regards,

    Wembanyama

  • 0 in reply to Wembanyama
    TI__Expert 7755 points

    Hey Wembanyama,

    Thank you for the additional details on your application. Based on that information, I would highly recommend looking through mmWave SDK Module Documentation (<MMWAVE_SDK_VER>/docs/mmwave_sdk_module_documentation.html), specifically the control layer and datapath sections. The range and AoA DPU sections may be particularly useful for your application as they outline how range, azimuth, and elevation is obtained from raw data. 

    I'm not sure why the links aren't working, but you should be able to access the following link https://www.ti.com/tool/download/RADAR-TOOLBOX/1.30.01.03 and select browse. This should take you to the TI Resource Explorer where you can find the MRR demo under Example Projects -> ADAS -> Medium Range Radar and the Gesture Recognition demo under Example Projects -> Gesture Recognition -> Gesture With Machine Learning. I have also attempted to relink them in this reply.

    Regards,

    Kristien