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.

  • Resolved

IWR1642BOOST: People Counter Application Narrow Range Issue (See Below)

Prodigy 80 points

Replies: 8

Views: 130

Part Number: IWR1642BOOST

Hi All,

I am currently using the TI IWR 1642 Boost in combination with TI's People Counter Application to track people moving through a room in our schools library. With the default configuration file the sensor would only reach 6 meters away, which did not cover the entire room. But after finding the "Office Space" configuration file, which increased the range to 15m, it worked perfectly. Correctly tracking my colleagues walking around the entire room. 

Here comes the issue,

Recently, within the last week, the gui for the people counter application stopped showing people farther than 6m on the point cloud but said that there were people still in the room. So like if someone walked around in a circle around 12m from the sensor, it would say that there is 1 person in the room and the points in S(:,n) correctly track the person. This is odd since it previously showed it in the gui.

This raised a flag that something else might have messed up too and sure enough, the sensor seems to have a much more narrow fov...

We would previously set up the sensor in a corner pointing at a 45 degree angle away from each wall covering a 90 degree range, so we could track the entire room and not miss any positions. However now it seems like it is much more narrow, losing targets if you walk around 25 degrees off in either direction of where the sensor is pointing. This happens for both the default config as well as the office space (extended range 15m) config.

Any help regarding this is much appreciated!

Just in case this is helpful here is the default cfg:

dfeDataOutputMode 1
channelCfg 15 3 0
adcCfg 2 1
adcbufCfg 0 1 1 1
profileCfg 0 77 30 7 62 0 0 60 1 128 2500 0 0 30
chirpCfg 0 0 0 0 0 0 0 1
chirpCfg 1 1 0 0 0 0 0 2
frameCfg 0 1 128 0 50 1 0
lowPower 0 0
guiMonitor 1 1 0 0
cfarCfg 6 4 4 4 4 16 16 4 4 50 62 0
doaCfg 600 1875 30 1
trackingCfg 1 2 250 20 200 50 90
sensorStart

And here is the office space cfg:

dfeDataOutputMode 1
channelCfg 15 3 0
adcCfg 2 1
adcbufCfg 0 1 1 1
profileCfg 0 77 30 7 62 0 0 24 1 128 2500 0 0 30
chirpCfg 0 0 0 0 0 0 0 1
chirpCfg 1 1 0 0 0 0 0 2
frameCfg 0 1 128 0 50 1 0
lowPower 0 1
guiMonitor 1 1 0 0
cfarCfg 6 4 4 0 0 16 16 4 4 50 62 0
doaCfg 600 1875 30 1
SceneryParam -1 1 0.05 15
GatingParam 4 3 2 0
StateParam 5 5 10 100 5
AllocationParam 100 0.01 8 1 2
VariationParam 0.289 0.289 1.0
trackingCfg 1 2 250 20 200 50 90
sensorStart

 

Thank you in advance!

  • Hello Matt,
    I've asked an mmWave Expert to look into this, we should have an answer for you in the next few days.
    Cheers,
    Akash
  • In reply to Akash Gondalia:

    Hi Matt,

    I want to confirm a few things:

    1. Have you made any modifications to the device firmware?
    2. Have you made any modifications to the configuration file?
    3. Have you made any modifications to the hardware?
    4. Have you made any modifications to the GUI?

    There are two problems:

    1. Point cloud does not appear in the GUI for distant targets (past 6 meters)
    2. Targets are only tracked in a +- 25 degree FOV.

    Regards,

    Justin

  • In reply to Justin Curewitz:

    Hi Justin,
    Thanks for helping me with this, below are the answers to your questions:

    1. No I have not modified any of the devices firmware. I flashed the mss and dss files onto the mmwave sensor months ago and have not touched the source code for them.

    2. No, the two configuration files that I attached at the bottom of my question above are exactly the same as when I got them. The default configuration is what came with the people counter application. I downloaded "Case-2, 14m (MIMO), Office Space" from dev.ti.com/.../node and that is what I have been using to track the entire room. My mistake it was 14m instead of 15m. Also, I have no clue what any of the values mean in these configuration files. Looking at what values are changed between the 6m and the 14m does not really make a whole lot of sense to me, even after looking at the documentation so I have not tried modifying it it all. If you have a better explanation for what the numbers correlate to that would be a life saver! :D 

    3. As for the hardware of the sensor I had purchased the devpack months ago thinking it would help me extract the coordinates of the people the mmwave sensor was tracking. I attached it to the IWR1642 when I received it but I never needed it. As I said before, The people counter application was working perfectly around 2 weeks ago, so I didn't see a need to remove the Devpack using the mindset, if it ain't broke don't fix it :). Do you think removing it might have an effect?

    4. I have made no changes to the GUI of the people counter application. After It stopped showing tracked people farther that 6m away I tweaked the boundaries a little with no success so I reverted back to the default.

    And yes those are my current issues.

    As for the MATLAB code, I have made a couple of changes so that I can extract the data tracked by the sensor. At the top of main I open a web socket server and connect to a separate program that will be receiving the data real-time. And inside the loop dealing with the tracked objects, I extract them from S(:,n), place them into a JSON string and send it over the socket connection.

    That is my end goal, which was previously working perfectly as I previously mentioned. When troubleshooting, I ran the original unedited MATLAB code that I made a copy of to save the original in case this exact scenario occurred. Same outcome, so I believe it does not have to do with my changes.

    Again thanks for your help on this,

    Matt

  • In reply to Matt Podzielinski:

    Hi Matt,

    Since there have been no changes to the software, we need to investigate the hardware. Please take the following pictures of the board:

    1. Front of board
    2. Side of board
    3. Close up picture of the antennas.

    Regards,

    Justin

  • In reply to Justin Curewitz:

  • In reply to Justin Curewitz:

     

    I am also getting these errors when loading up the people counter gui.exe from the quick start folder that was not there previously...

    Not sure if this is related or not 

  • In reply to Matt Podzielinski:

    Hi Matt,

    From the pictures, I am worried that the antenna on the board has been damaged. Please also try running without the dev pack attached, that should resolve the issue where the GUI fails.

    Regards,

    Justin

  • In reply to Justin Curewitz:

    Hi Justin,

    I had the idea to recompile the mss and dss files and re-flash them onto the sensor and that seems to have fixed the narrow range issue I was having. No idea why, it is very bizarre. Technically that is the main issue that was holding back my senior project team, the gui was just a clue that something else might be wrong. So since the range issue is fixed I do not want to change anything that may result in something going wrong again, so I'm under the impression, if it ain't broke don't fix it with the devpack right now. If another issue arises I will reopen this thread but for now I believe my issue is resolved.

    Thanks for your prompt responses Justin!

    Have a great day,

    Matt

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.