• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 264 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: CFAR calculation in mmw Demo for XWR68XX

Cagri Tanriover
Intellectual 905 points
Part Number: IWR6843ISK

Hi,

I am currently capturing data using the mmwave demo for IWR6843ISK and using the mmWave Demo Visualizer.

Following is the reference I am using for TLV packet decoding:

/ti/mmwave_sdk_03_05_00_04/packages/ti/demo/xwr68xx/mmw/docs/doxygen/html/index.html

Besides getting the detected objects with their x, y, x, velocity information (as per DPIF_PointCloudCartesian_t structure) I would also like to get the "Side information of the detected objects" of type MMWDEMO_OUTPUT_MSG_DETECTED_POINTS_SIDE_INFO

According to the DPIF_PointCloudSideInfo_t struct reference, following information is provided for each detected object:

Following are my questions:

1 - snr is an indication of how good the signal-to-noise ratio of a particular detected object is. Therefore, high snr can be used as a confidence of good detection for each object. Is this assumption true?

2 - noise is the measure of noise in the vicinity of the target object. Therefore, if I add the noise (in dB) to snr (in dB) I can calculate the signal level (in dB) for the detected object. Is this line of thinking true?

3 - What type of CFAR is used for the snr and noise calculations above? Is there a TI reference that describes this computation in detail?

Thank you for your help.

Regards...

 

  • 0
    TI__Expert 8802 points

    Hi,

    1. You are correct, a higher SNR indicates a stronger reflection.

    2. In theory, multiplying SNR * noise level would give a signal level for the detected point. This information isn't too useful, however. SNR is a better indicator of presence/strong reflections.

    3. Please check out the detection layer tuning guide for more information about our CFAR algorithms. https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.10.00.13&node=A__AIQPG9x7K34A8l4ZELgznA__radar_toolbox__1AslXXD__1.10.00.13 

    Regards,

    Tim

  • 0 in reply to Tim Henderson
    Intellectual 905 points

    Hi Tim,

    Thank you for the clarification. Those are really helpful.

    I have one more follow up question on this based on the data I have collected indoors. For a person sitting directly at boresight of radar at 260 cm, I am getting snr of 10 dB to 15 dB while the noise value is around 40 dB. Do these values sound reasonable?

    I convert the uint16 snr and noise values into dB by simply multiplying 0.1 dB with the uint16 value to get the above readings. In other words, if the uint16 value I get is say 240, the dB float value is 240 x 0.1 = 24 dB Is this calculation correct or am I missing anything important here?

    Regards...

  • +1 in reply to Cagri Tanriover
    TI__Expert 8802 points

    Hi,

    The values you are getting for snr and noise seem reasonable for an indoor environment. The snr and noise values depend on many factors, such as the distance, angle, reflection, absorption, interference, modulation, bandwidth, etc. of the target object and environment. Generally, having a higher SNR value is better. However, there is no definitive answer to what values are optimal or acceptable, as they may vary depending on the application and scenario. You may need to experiment with different settings and conditions to find the best values for your purpose.

    With regards to your calculation, that looks correct.

    Regards,

    Tim