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: Running a new Qt version

Part Number: IWR6843ISK-ODS

Attempting to run either fallDetection_main.py or gui_main.py on Raspberry Pi. I've installed the necessary packages/libraries, but it seems that pyqtgraph requires a Qt version 5.12 or higher. The Raspberry Pi comes loaded with version 5.11.3. I went through the lengthy process of building and installing Qt 5.15, but still getting a version error when attempting to run. Is there a step I'm missing to make pyqtgraph recognize the right Qt version? How would you recommend solving this problem? Thanks.

  • Hello Brian,

    We have not tested this GUI on the raspberry PI, although it should work. I have Pyqt5 v 5.15.2 installed and it works on windows. Python package support is outside of the support for these devices, but I would make sure your environment variables point to all the right versions. Can you verify with PIP or whichever package manager you use?

    Also, what are you using the GUI for on the raspberry PI? I worry the graphics may be too much for the PI to handle well and might run slowly. If you don't need the graph, you can use the parser and fall detection logic to make an interface without the pyqtgraph.

    Regards

    Jackson

  • Hello Jackson, 

    We are having a particular issue specifically with Qt recognizing the pyqtgraph not Pyqt5. The pyqtgraph is not recognizing the newly installed Qt version 5.15 on the Raspberry Pi it is still trying to recognize the 5.11.3 version.  

  • Hi Brian,

    Ok I understand. I am not sure how to help, I'm certainly not an expert in python packages. Can you uninstall the old version of the package? Not sure if the new one replaced or installed on top of 5.11. Otherwise, I am going to close this ticket as the python support is outside the scope of this forum.

    Regards,

    Jackson

  • Hi Jackson,

    We have been unable to get the Raspberry Pi to connect because of this issue. We are currently trying a Jetson Nano board. How would you recommend getting data out of the mmWave sensors for prototyping? Is there a TI solution? It seems to be enough of an issue that a company is making 3rd party hardware and software that integrates with the mmWave sensors because this issue is underdeveloped by TI. www.joybien.com/Buy.html

  • Hello Brian,

    There are a few other options to connect to the device. The python code we offer is just one example implementation

    The easiest is to run the evaluation on a windows PC. Is there a reason you need the R.Pi at this stage? I think the full visualized output will be rather heavy to run on an R.Pi or nano board. 

    The next step would be to take the example code from the visualizer to remove the visualization pieces and tweak the fall detection to fit your application. What does the end application implementation look like when running on the R.Pi? The parser scripts inside the oob_parser.py file can be used to parse the UART stream to get the data and perform similar fall detection logic on it.

    It does look like the python SDK offered in the above link is also based on pyqtgraph so maybe they have info on installing the correct packages, but I do not have the details of this offering.

    Regards,

    Jackson

  • Hello Jackson,

    The goad of the R.Pi was to conduct a small pilot without having to use a laptop for every instance. However, the difficulty in connecting the R.Pi and Jetson has resulted in buying small windows computers for the tests. 

    How do you recommend connecting to the cloud for devices at scale? If we were to do a pilot study with 50 IWR6843-ODS for fall detection and vitals, is there documentation on the best practice for hardware and software development?

  • Hi Brian, 

    There are many different approaches to connecting all these to the cloud. Wifi or other BLE style connectivity, ethernet connected with a R.Pi or beaglebone. It would all depend on your application needs. But this is not really dependent on the radar chip you are working with. I suggest looking at the wifi launchpad for connectivity options. Easiest way would probably be to send the UART data from the radar EVM to one of the wifi EVMs or something similar. 

    Regards,

    Jackson

  • Hello Jackson,

    Which Sensor-to-Cloud gateway solution would you recommend for a multiparameter patient monitoring solution? We would like to talk with someone about MCU architecture and engineering at some point. 

  • Hi Brian,

    This will depend too much on how you want to connect. Please have a look at the wireless connectivity devices at the link below. If you post a new question with one of those part numbers it will link the thread to the relevant experts, as this forum is just for radar parts.

    https://www.ti.com/wireless-connectivity/overview.html 

    Regards,

    Jackson