IWR1642BOOST: The numbre of detected objects drops dramatically at the location around (-0.2,4) in the ppclount demo with iwr1642

user5903056

Part Number: IWR1642BOOST
Other Parts Discussed in Thread: IWR1642, IWR6843

Hi,

I ran the pplcount demo with IWR1642 and I found that the number of detected objects drops dramatically at the lacation around (0.2, 4) which resulted in the break of tracking and the miscount of people number.

The below figure shows the problem. I would like to know the reason for this problem. It is said in TI's training document that "The angle estimation is most accurate at θ close to zero.", which seems conflit with the my observation.

Regards,

over 5 years ago

0 Nitin Sakhuja over 5 years ago

TI__Mastermind 30405 points

Hello,

A few suggestions:

1. Please ensure that the mounting angle (downward tilt) is setup according to the pictures shown in the demo user guide. Essentially, the if the mounting angle is too high or too low such that the targets fall at the edge or outside the elevation FoV (+/- 15 degrees), the SNR and hence angle measurement will be impacted due to elevation FoV even if the Azimuth angle is close to boresight.

2. While keeping your current setup, please try to run the Out of Box demo on your EVM and check if you see a loss in detection at the same region. Then make sure you follow the mounting consideration given in item#1 above and repeat the experiment. If you see loss in detection in any part of the FoV, that will translate to the people counting demo as well.

Regards

-Nitin

0 user5903056 over 5 years ago in reply to Nitin Sakhuja

Intellectual 910 points

Hi Nitin,

Thanks for the quick reply!

I will try it and give you the feedback as soon as possible.

Regards,

Jie

0 user5903056 over 5 years ago in reply to Nitin Sakhuja

Intellectual 910 points

Hi Nitin,

I've run the pplcount demo again as your suggestion. It seems there is a zone that has few detected points around boresight line, which is shown below. I think the reason has nothing to do with mounting angle.

Besides I've also flash the OOB fireware and observe the results on the Demo Visualizer, it is shown that OOB demo does not have this problem. Do you have some idea about the reason?

Regards,

Jie

0 user5903056 over 5 years ago in reply to Nitin Sakhuja

Intellectual 910 points

Hi Nitin,

Any response?

Regards,

0 Nitin Sakhuja over 5 years ago in reply to user5903056

TI__Mastermind 30405 points

Hi Jie,

Which version of the Toolbox are you using? What is the Silicon Revision of your EVM (Please refer to this page on how to find out the Silicon revision)

Regards

-Nitin

0 user5903056 over 5 years ago in reply to Nitin Sakhuja

Intellectual 910 points

Hi Nitin,

I've tried toolbox2.3.1 and toolbox 3.3.1. The silicon revision is ES2.0.

Regards,

Jie

0 Nitin Sakhuja over 4 years ago in reply to user5903056

TI__Mastermind 30405 points

Hi Jie,

1. The 1642 based people counting demo has been deprecated and provided as-is. We would recommend you to switch to the IWR6843 based people counting demo from the latest Industrial Toolbox.

2. If you still want to try the 1642 demo, I would suggest trying out the most recent version of this demo from toolbox 4.1.0. This was the last toolbox version that contained this demo before it was deprecated.

Regards

-Nitin

0 Liang Kun over 4 years ago in reply to Nitin Sakhuja

Prodigy 211 points

Hi Nitin,

I have tried the demo from toolbox 4.1.0, but the problem still exists.

Regards

Liang

0 Nitin Sakhuja over 4 years ago in reply to Liang Kun

TI__Mastermind 30405 points

Hi Liang,

1. Please try the demo on another EVM if possible.

2. As mentioned earlier, this demo is provided as-is and we recommend you to switch to the IWR6843 based people counting demo from the latest Industrial Toolbox.

Regards

-Nitin

Sensors

Sensors forum

IWR1642BOOST: The numbre of detected objects drops dramatically at the location around (-0.2,4) in the ppclount demo with iwr1642