• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 30 subscribers
  • Views 701 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

IWR6843ISK-ODS: Questions on demo firmware and interference

Nicholas Carley
Genius 9925 points
Part Number: IWR6843ISK-ODS
Other Parts Discussed in Thread: IWR6843,

Hello,

I got the below questions from a customer.

1.) TI 3D People counting demo firmware for IWR6843 is using Capon beamforming for point detection. Wonder why other TI firmware which return point clouds only does not use capon beamforming ? Is there any plan to add capon beamforming for normal point cloud firmware. Our understanding is that capon beamforming is more accuracy for point detection.

2.) The TI 3D people counting demo firmware seems to be designed for wall mount antenna. Does it support the IWR6843ISK-ODS ? Seems we can change antGeometry0/antGeometry1 for ODS antenna, anything we need to change to support that ?

3.) In order to cover a larger space for people counting, we might want to use two or more IWR6843 sensors, but we need full 4G bandwidth for full precision. So cannot do frequency division multiplex between devices. Since there are overlapped region between sensors to ensure coverage, what is the best way to avoid interference ?

4.) Does TI have some kind of detection when interference occurs ? 

  • 0
    TI__Mastermind 28055 points

    Hi Nicholas,

    1. Capon is more accurate, it is also much more CPU intensive and has a shorter max range. The OOB demo (point cloud) uses a simpler chain that is easier to configure and works better for long range detection (think 50+ meters for a human). If you want to use 3D People Counting like a point cloud demo, you could disable tracker in the software or even use the gui option to turn off tracker output.
    2. We provide a configuration for ODS as well.  Find it in the chirp_configs folder in the lab.
    3. Interference is very low probability if you are only using two devices. However, you could try the following:
      1. Binary phase modulation - this is a device level option that flips the phase 180 degrees randomly on chirps. This allows the device to "encode" the signal and make sure it gets the right one back. This isn't currently enabled on 68xx devices but will be in the future.
      2. Random delay between each chirp. Interference can only happen if the chirps are within the sampling frequency of eachother. Max sampling frequency on the device is 12.5 MHz, in 3D People Counting we are at 3 MHz. It is very unlikely for the devices to line up like this. By randomizing start time of each chirp (or frame) slightly, likelihood of the occurrence of this chirp overlap is decreased, and with the randomness, it wouldn't happen twice in a row.
    4. It is possible to detect when interference occurs, we haven't enabled this in the demos.

    Regards,

    Justin

  • 0 in reply to Justin Curewitz
    TI__Genius 9925 points

    Thank you Justin! I had some follow up questions:

    1.)    For the TI 3D people counting demo firmware using the ODS antenna. Does that firmware support ceiling mount option ? Seems the ODS board is designed for ceiling mounted, but all the documents  for the TI 3D people counting demo firmware shows that the device has to be wall mounted.   And what is the accuracy or coverage difference between the ODS and ISK antenna for the TI 3D people counting demo. Also what shall we modify if the TI 3D people counting demo does not support ceiling mount.

    2.)    In a production product, do we need the XDS chipset on every production device shipped, can the IWR works without that chipset ? Is there any magic the XDS chipset doing during the IWR chipset bootup and normal usage, or XDS is primary for debugging and signal muxing to PC during development only ?

    3.)    Does the mmwave firmware retrieve information from the EEPROM and behave differently based on the information in EEPROM. Or the EEPROM is only for housekeeping information only, such as serial number , sku, etc.

    4.)    How to retrieve the reading from temperature sensor and current sensor on the ODS board ? Does the mmwave  firmware currently implemented  or use that ?

    5.)    For IWR6843 HS device, are we able to burn the public key ourselves or we need to provide public key to TI, and TI will burn it ?

    Thanks,
    Nick

  • 0 in reply to Nicholas Carley
    TI__Mastermind 28055 points

    Hi Nick,

    1. The firmware currently available is intended for wall mount.
    2. No - XDS110 is for debug and flashing the EVM, it is recommended that you do not include it during production.
    3. It is really for housekeeping. The device type is defined in the soc driver.
    4. See this post: https://e2e.ti.com/support/sensors/f/1023/t/790909?IWR6843-Could-not-read-all-internal-device-temperature-values-
    5. You will burn the public key yourself.

    Regards,

    Justin

  • 0 in reply to Justin Curewitz
    TI__Genius 9925 points

    Hi Justin,

    Following up with #1. 

    What is the accuracy or coverage difference between the ODS and ISK antenna for the TI 3D people counting demo? Also what should we modify if the TI 3D people counting demo does not support ceiling mount?

    Thanks,
    Nick

  • 0 in reply to Nicholas Carley
    TI__Mastermind 28055 points

    Hi Nick,

    ODS covers a 120x120 degree FOV, but has lower gain resulting in lower max range the the ISK. ISK covers a 120 x 40 degree area. ODS has better elevation resolution and accuracy than the ISK, but worse azimuth resolution and azimuth accuracy than the ISK.

    Is ceiling mount supported? There is nothing stopping you from running the device on the ceiling and detecting people. In this case, the ODS is preferred for the larger FOV. Performance may be good enough for your use case. The software was not designed with ceiling mount in mind, nor has it been tested for performance when mounted on the ceiling.

    Regards,

    Justin