Other Parts Discussed in Thread: IWR6843ISK
Hello,
We are trying to enable 3D People Counting for Long Range case. I refer to the Max Range and Range Resolution formula (from GUI source code) in modifying the slope constant and sampling rate in profileCfg line in order to get my desired range.
Result:
- For our ISK board used with Sense and Direct firmware, the 10m cfg that I modified works fine. We are getting good detection up to 10m. This is 2D though.
- However, for ODS board used with 3D People Counting firmware, the response is not that good - the point cloud at 8m and beyond is too few. For this, we tried to adjust the cFarcfg particularly the rangeThre and angleThre. If we lower it too much to get enough point cloud at longer range, the downside is it will scatter too much. We tried different combinations for rangeThre and angleThre. We will get enough point cloud (but result to scattering) at rangeThre = 2.5 / angleThre = 2.5. See attached pic for point cloud of 1 person at around 8m. If you are interested in the whole video, I attached it as well (play with VLC player).
profileCfg 0 60.75 7 7 41.10 0 0 38.71 1 96 2950.00 0 0 24
Questions:
- Is it a good idea for ODS - 3D PC to be used in Long Range (10m)? Or the antenna design only allows it to work for 8m and below just like the default demo. If it can work for 10m and above, please help on what we might have missed in enabling longer range.
- Why is the point cloud scattering for 3D ODS and not for 2D ISK at 8 to 10m range? Note that we used the same profileCfg config. I believe the 3D ISK also suffers the same point cloud scattering but I have to double check.
