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.

MMWAVE-3P-SEARCH: Coherence of calculations between mmwave sensing estimator and sample config files

Part Number: MMWAVE-3P-SEARCH

Hi,

  I am working with mmwave 6843 ISK from TI. I plan to modify the config files provided by TI to meet my application scenario (range/resolution/velocity etc).

I am planning to use mmwave sensing estimator (https://dev.ti.com/gallery/view/mmwave/mmWaveSensingEstimator/ver/1.4.0/) as a tool for setting my chirp parameters.

However, to start with, I expect that I must be able to regenerate the same config parameters as are in a sample config file (say ISK_6m_default). Is there a document that explains how the numbers in a sample config file (say ISK_6m_default) have been arrived at, considering the mmwave sensing estimator as reference. Alternately, if there is a document explaining the numbers in the config file based on calculations mentioned in mmwave fundamental document here, that is also ok.

Regards,

  • Hi,

    Firstly, I would like to make it clear that the sensing estimator shows what is possible via the radar front end on the device, but when running full labs/demos, there may be additional restrictions which would not be reflected in the sensing estimator. For example, the sensing estimator will assume that you can utilize the full radar cube size, but some demos may need to use some of this space for application data storage. 

    In regards to how the numbers are arrived at for the configurations included with the demos in the mmWave Industrial Toolbox, this is typically done by tuning the configuration file for what worked best when developing the application. These may not be the most ideal parameters for every demo for every environment. For example, if you are running the overhead 3D People Counting demo, this configuration will likely need to make some assumptions related to the height of the ceiling which the device is mounted to. If you were then to take this configuration, and run it on a sensor which is 30 meters off the ground, it would not perform the same (an extreme example, but just illustrating the point. 

    Please keep in mind that the demos/labs in the mmWave Industrial Toolbox are intended to serve as a starting point for your development, and may not give you the most optimal results as is. The best advice I can give is to take a configuration file for a demo you are interested in, input it into the sensing estimator, and take a moment to try to understand the tradeoffs which were made (such as memory usage, framerate, doppler resolution, range resolution, max range, etc) and try to understand why they may have been chosen for that application. Once you feel that you have established this understanding, you can then make adjustments to suit what is needed for your use case.

    Let me know if you have any further questions.

    Best Regards,
    Alec

  • Hi,

       Firstly, thanks for the prompt response.

    From the response, my understanding is that the sensing estimator provides numbers based on theoretical calculations (more of the best case), whereas the config files provided with the labs are derived based on practical use cases and practical device limitations. Thus, taking the lab config files as a starting point for implementation makes more sense. Is my understanding correct?

    In this case, some explanation of how the values in config file are arrived at, to aim at a particular usecase (range, range resolution, velocity, etc) becomes all the more important. Unless this is provided, it becomes difficult to adapt these config files.

    Let me illustrate my point through a simple example. The people counting user guide here provides certain values in a table in Step-4. For a bandwidth (B) of 2249 MHz, the range resolution stated is 0.084m. My calculation (c/2B) gives me a number of 0.066m. I do not know whether the difference is due to any miscalculation on my part or any practical approximation from TI.  I can tweak the parameters only after some reasoning for their current values.

    In general, I have liked the exhaustive documentation provided by TI on mmwave. Especially the introductory videos, labs and explanation of parameters . Some more insight into how the config files have been written would be helpful. 

    Regards

    Girish

  • Hi Grish,

    First, "Thus, taking the lab config files as a starting point for implementation makes more sense. Is my understanding correct?": Yes, if what your application is is aligned with one our of application demos in the industrial toolbox then one of the configurations included with application demo is a great spot to start. The equations in the sensing estimator are derived from this application note and can be used to look at a configuration and derive max range, range resolution, max velocity and velocity resolution. 

    Now the sensing estimator and the calculations in this app note are just approximations for the scene parameters (range resolution, velocity resolution, max range and max velocity). In our mmwave studio tool you will find the ramp timing calculator which takes into account of digital front end delays. These delays can effect the bandwidth and hence effect the range resolution. 

    The explanation of parameters you linked is a great place to start for understanding the parameters.

    Thanks,

    Angie