• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 30 subscribers
  • Views 432 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-ODS: Default values of calibration parameters and how does calibration affects pointcloud data

Victoria Rudakova
Intellectual 770 points
Part Number: IWR6843ISK-ODS

Hello TI team,

I've been looking at the Overhead 3D people counting lab, and noticed one of its requirements is the calibration procedure in order to obtain compRangeBiasAndRxChanPhase coefficients.

I also noticed that the default values of those coefficients are "0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0", while the "3D point cloud data" lab has default coefficients as "0 -1 0 1 0 1 0 -1 0 -1 0 1 0 1 0 -1 0 -1 0 1 0 1 0 -1 0" (some of them are negative, and as I understand it is so that to take into account rotations of some of the antennas). 

1. So how come the default coefficients of the compRangeBiasAndRxChanPhase parameter are different for the two firmware? When I tried to use the default values with negative signs, the point cloud coordinates got inverted along X axis, it seemed. Could it be the "overhead people counting" firmware negates the necessary coefficients on the firmware level?

I had also performed a calibration procedure with my sensor and obtain the calibrated values so that to use them within the "overhead people counting" lab. 

2. Is there any visual difference that is made by using the calibrated values? Is it something that I'm expected to see? How does that difference help the "overhead people counting" lab to be more precise? I'm just trying to understand if the calibration procedure is critical and whether it has a great impact on the performance of the "overhead people counting" lab.

Thank you.

Best regards,

Victoria

  • 0
    TI__Mastermind 30405 points

    HI Victoria,

    1. The default coefficients are the same for both ODS and AOP in the overhead demo, just like the OOB demo from where this command is borrowed. However, there is an antPhaseRot command in the CLI which specifies the phase rotations to be applied when using the default coefficients. Note that the antenna specific phase rotations are required / used only for default coefficients.

    When you measure the coefficients using the calibration procedure, the measured values already include the phase information (i.e. are pre-rotated according to the antenna pattern) and hence those should not be inverted again. So, the overhead demo code code checks the values and if it finds the default values, it applies the antPhaseRot command to invert those according to the antenna pattern. (left is ODS, right is AOP)

       
    Also, note that the pre-requisite mentions which file to use as base according to your EVM, because the antenna geometry and phase rotations are specified accordingly.

    2. We have not quantified the effect of not calibrating the coefficients for your EVM but the general impact is listed in the pre-requisite section as shown below. Basically, there will be range errors due to range bias and velocity and angle measurements errors due to uncalibrated phase measurements. All these will add to overall tracking inaccuracies.

     

    Regards

    -Nitin

  • 0 in reply to Nitin Sakhuja
    Intellectual 770 points

    Hi Nitin,

    Thank you for the detailed answer, I appreciate it.

    Regards,

    Victoria