IWR6843ISK: Effect of configuration files on diffuse reflection

Additionally, is there a way to verify the config made by the "mmWave Sensing Estimator"?

KIM TAEYOUNG said:

Using the 70m config provided by the "traffic monitoring lab", there is almost no reflex.

Are you saying you are not getting the amount of points you were expecting or are you seeing more points than what you were expecting?  If it is the former, then you can try increasing your RxGain.  If it is the latter, you can try increasing your CFAR threshold to eliminate noisier points with lower SNR.  In addition, after plugging your configuration into the sensing estimator, it looks like your ramp end time is not valid, so I would adjust that as well.

Sincerely,

Santosh

There is a phenomenon in which points of an object do not gather and spread sideways. In addition, the snr values of the points spreading sideways are actually not much different from the snr values we want, so if you raise the cfar threshold, they are not captured as objects.The ramp end time has been readjusted.

Regards,

Kim

Hello.

Are you seeing this behavior with both the 70m config in the traffic monitoring lab and your custom config or is this only seen in the custom config from sensing estimator?  And can you send your updated configuration?

Sincerely,

Santosh

Hello

This phenomenon is only visible in custom configurations. Switched to rampend time 21.74 -> 25.15.

Regards,

Kim

Hello.

What exact requirements are you trying to achieve with your custom configuration?  I would recommend you use the 70 m configuration as a starting point and modify it to meet your requirements.  I noticed that your start frequency is much higher and your frequency slope is much lower; if you are looking to boost your maximum range, I would start with just adjusting the frequency slope.

Sincerely,

Santosh

In terms of user requirements, the goal we want to achieve is to detect the vehicle about 150m. However, increasing the distance increases diffuse reflection and noise, which causes problems in object detection. When we tested it ourselves, we measured and set the longest distance when the value of the start frequency was 63 to 63.5. Is there any other good value?
Regards,
Kim

Hello.

Your configuration should still work at 60 GHz, which is what the 70m configuration is set at.  Changing the frequency slope should be enough to reach your range requirements.

Sincerely,

Santosh

Hello

If so, if the starting frequency value is set to 60, what is the appropriate value of the frequency slope when you want to detect 150m?

Regards,

Kim

Hello

Should I find a frequency gradient that fits 60ghz and apply it? I can't think of a good idea.

Best Regards,

Kim

Hello.

You can start with just modifying the frequency slope in the sensing estimator to meet the distance requirement, and once you send the config, if it errors then you know that 60 GHz + that frequency slope will not work, and you can adjust it from there.  But I wouldn't expect it to fail unless the frequency slope goes extremely low.

Sincerely,

Santosh

Using the sensing estimator, the following values were obtained.
profileCfg 0 60 2 3.8 21.68 0 0 9.95 1 210 12441 0 0 30
channelCfg 15 7 0
adcbufCfg -1 0 1 1 1
chirpCfg 0 0 0 0 0 0 0 1
chirpCfg 1 1 0 0 0 0 0 2
chirpCfg 2 2 0 0 0 0 0 4
frameCfg 0 2 64 0 40 1 0

Is there any more price to change here?

Regards,

Kim

Hello.

Have you ran this configuration with the demo to see if meets your requirement?  It is important to test it with the demo to make sure it doesn't go past any firmware limitations as this is all theoretical calculations.  Feel free to ping this thread with your results.

Sincerely,

Santosh