AWRL6432: Antenna coupling in AWRL6432 is not linear/constant across frames

Hi Girish,

Thanks for posting. I wanted to post a few more details not captured here to make sure our understanding is correct. Could you please confirm or elaborate on the bullets below?

• You have removed static clutter removal from your algorithm and you are implementing this because peaks near the sensor are being detected more with this change
• The antenna coupling suppresses the power of the near range bins for about 10 seconds before this behavior is occurs
• When you observe the radar cube/the raw data from the ADC, you are not seeing this behavior (i.e. the power does not increase for the near range bins after ~10 seconds)
• You are taking these measurements at boresight with no objects present for about 2 meters in an anechoic chamber 

Thanks again,

Clinton

Hi Girish,

Following up on this. Here are some additional questions/comments:

1. The behavior you are showing in these images is expected, given that you are not performing clutter removal. Did you perform these same tests on the previous design and not observe the power increase over time for the near bins?
2. Is there a reason you are not performing clutter removal? We would recommend including this.
3. We typically discard the first several bins when performing CFAR to remove this antenna coupling effect. Could you operate this way as well? If not, could you describe what you are trying to detect so close to the sensor?

Thanks,

Clinton

Hi Clinton,

Please find my answers below:

Clinton Rodgers said:

You have removed static clutter removal from your algorithm and you are implementing this because peaks near the sensor are being detected more with this change

Yes, We disabled Static clutter removal algorithm in order to see whether antenna coupling is really removing these high peaks.

Clinton Rodgers said:

The antenna coupling suppresses the power of the near range bins for about 10 seconds before this behavior is occurs

For some frames(not exactly 10secs) we saw the peak less than CFAR thresholds

Clinton Rodgers said:

When you observe the radar cube/the raw data from the ADC, you are not seeing this behavior (i.e. the power does not increase for the near range bins after ~10 seconds)

We observed only ADC data of the last chirp that power is almost same for 1st frame's last chirp and 800th frame last chirp. We have not looked into radarCube and whole ADC data.

Clinton Rodgers said:

You are taking these measurements at boresight with no objects present for about 2 meters in an anechoic chamber 

Yes, you are correct.

Clinton Rodgers said:

The behavior you are showing in these images is expected, given that you are not performing clutter removal. Did you perform these same tests on the previous design and not observe the power increase over time for the near bins?

We never observed this behavior in previous SDK like 3.5,3.6. In 6443 & 6843, when static clutter is disabled we have not seen any near points after applying antenna coupling.

Clinton Rodgers said:

Is there a reason you are not performing clutter removal? We would recommend including this.

Customer requirement is that they need to detect static objects without near points. They verified this setting(static clutter removal disabled and antenna coupling enabled) on 6443 and it is absolutely working fine

Clinton Rodgers said:

We typically discard the first several bins when performing CFAR to remove this antenna coupling effect. Could you operate this way as well? If not, could you describe what you are trying to detect so close to the sensor?

Yes, But as a customer requirement we need to detect any obstacles that are blocking the sensor, so we need to detect points close to sensor.

Regards

M Girish Veerendra

Hi Girish,

Thanks for the follow up. We'll wait to provide more feedback once you are able to perform the additional testing discussed.

Thanks again,

Clinton