• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 31 subscribers
  • Views 4743 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

IWR1443BOOST: Reading from serial on a raspberry pi

Dexter Banham
Prodigy 70 points
Part Number: IWR1443BOOST
Other Parts Discussed in Thread: IWR1443

Hello, 

I have written some python code that reads data from the IWR1443 BOOST EVM and interprets it (while the mmwave demo is running),

I have been wondering if I can set this up on a Linux machine, specifically a small arm based Linux machine. The issue I have been having is that without the emupack the serial ports cannot be found and the version of this software for Linux seems to be precompiled for an x86-64 Linux setup. 

My main aim is to get data from this EVM to a raspberry pi or other board so that it can be embedded in a mobile platform for a prototype robotics project, is this at all possible? 

Could the IWR1443 DEVPACK be used to this end at all? 

Thank you for your answers,

Dexter  

  • 0
    TI__Mastermind 21295 points

    Dexter,

    Thanks for your interest in the IWR1443 mmWave sensor device from TI.  I would be happy to assist you.

    There are a couple of ways to get the detected object data from an IWR1443 BOOST EVM running the out-of-box demo firmware to a microcontroller like the Raspberry Pi.  The easiest way is to use the standard UART over USB path that you have mentioned.  When connecting the EVM to a Linux machine (PC-based or ARM-based processor like Raspberry Pi) it should show up automatically as serial ports /dev/ttyACM0 and /dev/ttyACM1.  This relies on the CDC-ACM Linux module which is typically present/enabled by default in Linux.

    A second path which would take a bit more effort but would eliminate the dependency on the USB interface is to connect the EVM to the microcontroller directly over UART or SPI.  The UART interface is already supported by the EVM out-of-box demo firmware whereas the firmware would need to be modified (using the mmWave SDK) to use SPI instead.

    Please mark as Answered if this resolves the issue or reply if more support is required.

    Regards,
    John

  • 0 in reply to JohnH
    Prodigy 70 points
    Dear John,

    Thank you for your response, the solutions you have outlined seem very promising, however, I would like to ask a few follow-up questions about the details.

    Just to clarify, I am interested in the Range Profile and Doppler Range Profile rather than the detected objects per se.

    My current understanding is that to interface with the IWR1443 EVM over USB-serial the SDK for the mmWave chipsets and the emupack have to be installed on the host machine, can I get a confirmation that I can interface with the EVM over USB-serial without the two aforementioned software packages on the Raspberry pi? They seemed to me to be precompiled as .bin files and I could not run them on raspien OS on a raspberry pi.

    My second follow-up question is in regards to the DEVPACK for the IWR1443 EVM, would this board come in handy at all for my prototype? Or will I be able to connect the EVM to the Raspberry pi without it?

    Thank you in advance for your response,
    Best regards,
    Dexter Banham
  • 0
    TI__Mastermind 21295 points

    Dexter,

    I happened to specifically mention detected object data in my initial reply, but the same applies to any data type that is output by the out-of-box demo firmware (including Range Profile, etc.).

    On a Windows PC, the emupack drivers are required to communicate with the EVM. These drivers can either be installed in a stand-alone manner or they come as part of Code Composer Studio. The mmWave SDK is only required if you wish to modify/build the mmWave sensor firmware. (The SDK also comes with the pre-compiled binaries for the out-of-box demo firmware in case you ever need to re-flash the EVM.)

    On a Linux machine, the CDC-ACM Linux module is typically already present/enabled and will automatically allow you to access the two EVM serial ports at /dev/ttyACM0 and /dev/ttyACM1 (assuming those ports are not already in use by some other device) after the EVM is powered-up and connected to the Linux machine via a USB cable.  This allows communication via UART over USB with the EVM without any other software installation.

    However, if you wish to build or debug TI mmWave sensor firmware on the Linux machine (typically done on a PC, not an ARM processor), you would also need to install the mmWave SDK as well as Code Composer Studio.

    The DEVPACK board is only required if you wish to access more device pins than are accessible on the normal EVM or if you wish to capture the raw ADC Radar data for custom FMCW Radar processing algorithm development. (Capturing the raw data also requires a TSW1400 board connected to the DEVPACK.)

    If this resolves the question, please mark as Answered or reply if more support is required.

    Regards,
    John

  • 0 in reply to JohnH
    TI__Mastermind 21295 points
    Dexter,

    Just checking if any more support is required or if the question is resolved would it be possible to mark as Answered?

    Thanks,
    John
  • 0 in reply to JohnH
    TI__Mastermind 21295 points
    Closing this thread since there is no more activity. Please open a new thread if more support is required in the future.

    Regards,
    John