• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 387 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.

IWR6843AOPEVM: recognize and display static movements

Brian Zoro
Expert 4030 points

Part Number: IWR6843AOPEVM
Other Parts Discussed in Thread: IWR6843AOP, UNIFLASH,

Dear all,


Our customer ordered the IWR6843AoP EVM and did some first testing.
After first attempts with the delivered SDK version 3.4, they updated to 3.5. (flashed with Uniflash)
So the SDK version is now: 03.05.00.04

The hardware version: IWR6843AOPEVM ES02.00 (Rev. F)
The demos run on the PC with mmWave_Demo_Visualizer / ver / 3.5.0

In these demos, dynamic movements are shown or registered as standard.
If they stay “static” in front of it, the measured values “disappear” again. Somehow everything is tuned to Doppler radar.

However, they would need a setting where slow or almost and completely static movements are recognized and displayed.
Could you please give me suitable settings or a more suitable demo program than this “out of the box” demo?

The measuring range would be roughly: up to 2 meters, azimuth +/- 30 °, elevation: +/- 60 °

And one more general question about these radar sensors. Is it possible that people, bicycles, wheelchairs and prams can be recognized and differentiated with it? Is there a corresponding demo for this?

Thanks in advance.
Regards,
Brian

  • 0
    TI__Mastermind 23725 points

    Hello

    Have you tried the Static clutter removal feature which is part of the SDK OOB.  - Please see user's guide

    For detection of various objects can you please check the following app note:  mmWave Radar Sensors: Object Versus Range

    This will allow you to see how objects are detected. Classification is very different than detection and you may have  to look for specific attributes  based on object type.

    Please let us know what the results of static clutter removal show you.

    Thank you,

    Vaibhav

  • 0 in reply to Vaibhav Mahimkar
    TI__Expert 4030 points

    Hi Vaibhav,

    here the feedback from customer:

    I tried this "Remove Static Clutter" feature last week ...
    With this I can see the distance of an object better. (Somehow the "noise floor" is factored out ...) However,
    the measurement peak also disappears here if I keep the measurement object still at a fixed distance ...

    Your "mmWave Radar Sensors: Object Versus Range" actually shows that even with your more complex sensors "only" object detection with distance and / or speed measurement is possible, if I understand it correctly. It is not really possible to classify which object it is.

    For the pure radar detection (with distance measurement) of an object there seem to be simpler sensors that do not work with 3 transmitting and 4 receiving antennas.

    Kindly help me provide feedback on that.

    Regards,
    Brian

     

  • 0 in reply to Brian Zoro
    TI__Mastermind 23725 points

    Hello

    Please see inline:

    Brian Zoro said:

    I tried this "Remove Static Clutter" feature last week ...
    With this I can see the distance of an object better. (Somehow the "noise floor" is factored out ...) However,
    the measurement peak also disappears here if I keep the measurement object still at a fixed distance ...

    This feature allows  ignoring objects with 0 Doppler. so If the object is stead it will be eliminated.

    If the object has motion it will be retailed and passed through detection chain  to be compared against CFAR thresholds.

    It would be good for you what your range plots show with/without static removal.

    Brian Zoro said:

    Your "mmWave Radar Sensors: Object Versus Range" actually shows that even with your more complex sensors "only" object detection with distance and / or speed measurement is possible, if I understand it correctly. It is not really possible to classify which object it is.

    CFAR works agnostic to motion or static peaks.  In current implementation both the static as well as doppler based peasks are compared for object.  You may choose the mmwave data characteristics  to make judgment calls about type of objects if those let you tell them apart.

    Please see mmave people detection demo with classification and gesture demo to get an understanding for this.

    Brian Zoro said:

    For the pure radar detection (with distance measurement) of an object there seem to be simpler sensors that do not work with 3 transmitting and 4 receiving antennas.

    We are not sure what you mean by this. Please see the comparison table on the product pages or any of the datasheets for device level feature distinction.

    Thank you,

    vaibhav

  • 0 in reply to Vaibhav Mahimkar
    TI__Expert 4030 points

    Hello Vaibhav,

    Thanks for the feedback.

    ".....Please see mmave people detection demo with classification and gesture demo to get an understanding for this."

    could you kindly share the link to these demos?

    Regards,
    Brian

  • 0 in reply to Brian Zoro
    TI__Mastermind 23725 points

    Hello

    It has been  name and targeted for HVAC applications

    https://dev.ti.com/tirex/explore/node?node=AErmTL1DcBgtfS0zrdmYHg__VLyFKFf__LATEST

    It distinguishes between the motion from human (colored) vs motion from leaves/fan(grey).

    This demo conveys how mmwave data can be distinct for different kind of object/motions and a pattern can be established. 

    Thank you,

    Vaibhav