• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Subscribers 30 subscribers
  • Views 510 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.

IWR6843: SNR dependency on velocity

Pavel Bartoš
Prodigy 70 points
Part Number: IWR6843
Other Parts Discussed in Thread: SYSBIOS, MMWAVE-SDK

Tool/software:

When assessing our Doppler radar sensor, which utilizes the IWR6843 chip, we observed that the Signal-to-Noise Ratio (SNR) varies with velocity.

According to our findings, there appears to be a correlation where an increase of velocity by 3 m/s corresponds to a decrease of SNR by 1dB. We're unsure if this phenomenon is widely recognized as a standard behavior in radar systems, as we haven't found relevant literature discussing it.

We're curious about the underlying reasons behind this relationship. If you have any insights or suggestions, we'd greatly appreciate them.

  • 0
    TI__Mastermind 21071 points

    Hello, 

    Can you please share more about these findings? As velocity continues to increase do you see a continued decrease in SNR? Was this testing performed in an anechoic chamber? How are you generating a specific velocity?

    Best Regards,

    Josh

  • 0 in reply to Josh Dye
    Prodigy 70 points

    Sure, there are a few measurements we have conducted:

       

    Our tests were car tests, where we have attached the doppler sensor approximately 1m above the ground.
    Velocity is generated by the car movement.

  • 0 in reply to Pavel Bartoš
    TI__Mastermind 21071 points

    Hi, 

    Thank you. Can you please clarify further?

    Is the sensor attached to the car? If so is the sensor pointed directly at the ground or is the sensor pointed outward?

    Or do you mean the sensor is mounted in a fixed position and a car drives in front of it?

    Thanks,

    Josh 

  • 0 in reply to Josh Dye
    Prodigy 70 points

    The sensor is attached to the car and is pointing to the ground at approx. 30° angle.

    Profile config parameters used:

    Parameter

    Value

    noiseGain

    1.04

    noiseGainLimit

    1.04

    rxChannelEn

    1

    txChannelEn

    1

    startFrequency

    60

    idleTime

    15

    rampEndTime

    47.5

    txPowerOut

    0

    freqSlopeConst

    80.61

    digOutSampleRate

    7014

    framePeriodicity

    100

    frameTriggerDelay

    1

    satPrimarySliceDuration

    5

    satNumSlices

    121

    numAdcSamples

    256

    numChirpLoop

    512
  • 0 in reply to Pavel Bartoš
    TI__Mastermind 21071 points

    Hello, 

    Thank you for the diagram and additional info. So just to confirm, as the velocity increases, the distance between the radar and the ground remains the same, correct? Also, in the plots that you have shown above, is it true that they are including data which is compiled over a long time (multiple frames) or they are representing the SNR vs Velocity of every point in a specific frame?

    Best Regards,

    Josh

  • 0 in reply to Josh Dye
    Prodigy 70 points

    Hi,

    yes, the distance between the radar and the ground remains the same.

    Each point represents single radar measurement. Each plot represents different CAR test.

  • 0 in reply to Pavel Bartoš
    TI__Mastermind 21071 points

    Hello, 

    Thank you. I apologize for asking so many questions but can you also confirm which example software you are running? I'm discussing internally to determine if this phenomenon would be expected. Please give me a day or two to get back to you. 

    Best Regards,

    Josh

  • 0 in reply to Josh Dye
    Prodigy 70 points

    Hi, 

    we utilize your demo SW for IWR6843 chip: SYSBIOS:6.73.01.01;com.ti.MMWAVE_SDK:3.02.00.04;,xdcToolsVersion=3.50.08.24_core 60 GHz

  • 0 in reply to Pavel Bartoš
    TI__Mastermind 21071 points

    Hi, 

    Thank you for the info. Please expect an update tomorrow. Thanks for your patience. 

    Best Regards,

    Josh

  • +1 in reply to Josh Dye
    TI__Mastermind 21071 points

    Hi Pavel, 

    My sincere apologies for the delay. Is there a specific reason that you do not use the latest version of the MMWAVE-SDK (3.6)? I'd be curious to know if you observe the same phenomenon in your tests with the updated version. 

    Regardless, one potential explanation could be ‘range migration’. The radar measures a scene in finite time – for example a 15 ms active frame-time (with 128 equally spaced chirps) could be  used to sample a scene with a moving target.

    During that 15ms the target moves by a certain distance (‘range migration’), which would reduce its SNR compared to how it would be if the target was standing still. In some cases the targets velocity could be so high that it moves from one range-bin to another.

    There should be ways to compensate for this… you could estimate the targets velocity and then re-process the data after correcting for the targets velocity. At the moment, none of our example demos implement this compensation but it is something you could explore if you wish.

    Best regards,

    Josh