• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 539 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.

IWR6843ISK: Z-Axis Position of Points and Point Detection Question

Bill Li
Prodigy 20 points
Part Number: IWR6843ISK
Other Parts Discussed in Thread: IWR6843

Note; We are using this ROS Driver: github.com/.../src

We have two questions:

1. We were wondering whether it'd be possible to get Z-axis position of points, or is that always constrained to the z=0 plane? The data sheet specifies that the radar has an elevation detection range of 30 degrees. However, when viewing the point cloud in our simulation(Rviz), all the points lie on a 2D plane within the 120 degree azimuth detection range.

2. What determines whether a point is read? It seems when velocity is equal to 0, no points are recorded, when there is apparent velocity, what determines what points are read? (closest, highest density, random)?

Your response is much appreciated!

  • 0
    TI__Mastermind 23177 points

    Hello,

    If you look at the file "DataHandlerClass.cpp" in the ROS driver repo you linked, go to line 446 (395 if using old SDK which I do not recommend). This line is getting the Z-axis coordinate of a point via UART straight from the sensor. This means we are indeed have the ability to capture the Z-axis of a point. Now next look at line 835. We have the ability and indeed pull the Z-axis value of every point, but for the purpose of our demos such as Safety Bubble we only care about a 2D plane and so we set every visualized point to z=0. One of the steps will be changing the line marker.pose.position.z = 0; to marker.pose.position.z = msg.z; in order to send elevation data as well.

    Best Regards,

    Pedrhom Nafisi

  • 0 in reply to Pedrhom Nafisi
    Prodigy 20 points

    Hi Pedrhom,

    Thank you! We will look into it. Do you mind helping us with our 2nd question as well.

    Cheers,

    Bill

  • 0 in reply to Pedrhom Nafisi
    Prodigy 20 points

    Hi  Pedrhom,

    we noticed this several weeks ago and had made the change already, but the result didn't change. The note that you pointed to specifically is only for visualization marker too, so changing that line shouldn't affect the point cloud's z-axis values, which are also reporting 0.

  • 0 in reply to Bill Li
    TI__Mastermind 23177 points

    Hello Bill,

    For your second question: In the same file, you can find the calculation and algorithm on whether a point is kept or not starting on line 617. The calculation involves elevation and azimuth angles. Do note that the way this is all working is that the sensor is outputting all points, sending it over via UART, and that this DataHandlerClass.cpp has several areas like the area mentioned above that trim down what is sent to be published by RVIZ. It is here you can add/remove constraints on the points. If you wanted to, you are able to remove all constraints and send over every point, but it would include a lot of inaccurate/reflected points. 

    So you are saying the point cloud itself is saying that all points have a z-axis of 0, understood. What firmware is currently being flashed to the IWR6843 sensor? Is it the most recent version of the Out of Box demo, the binary provided with the SDK?

    Best Regards,

    Pedrhom Nafisi

  • 0 in reply to Pedrhom Nafisi
    Prodigy 20 points
    I'm not 100% sure which version is flashed. I know I re-flashed it around January, but I'm not sure which version that was. Is the most recent out of the box demo version the one that he would recommend?
    12:15
    Good to know about the constraints. It seemed like most of the constraints had to do with the elevation and azimuth inclusions, but perhaps I need to have a closer look
  • 0 in reply to Bill Li
    TI__Mastermind 23177 points

    The most recent version of the Out of Box demo binary can be found with our SDK and can be downloaded here: https://www.ti.com/tool/MMWAVE-SDK

    Another question: since this seems to be an issue sensor side and not ROS side, which cfg is being given to the sensor when ran? Note it must be the 6843ISK_3d.cfg and not a 2D one. I have attached it below for your convenience.

    Lastly, the git you linked seems very old, you can grab the newest version of the ROS driver by doing: 

    git clone https://git.ti.com/git/mmwave_radar/mmwave_ti_ros.git

    6843ISK_3d.cfg

    Best Regards,

    Pedrhom Nafisi