• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 2 answers
  • Subscribers 54 subscribers
  • Views 1745 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.

RTOS: CC2640: AoA algorithm Limitations

GOHM ELECTRONICS
Expert 2085 points

Tool/software: TI-RTOS

Hello, 

We want to use Localization toolbox in CC2640R2 SDK 2_20_00_49. I have few questions about limitations and accuracy of ToF and AoA algorithms.

What is the location update resolution (in seconds or milliseconds) for the ToF and AoA?

What is the accuracy (in degree for AoA, in meter for ToF) of ToF and AoA?

Best regards.

  • 0
    TI__Guru 56855 points
    Hi,

    The location update resolution for ToF depends on the duration of your burst. The more samples you have in your burst the better accuracy it will be, however, that also means you need more RAM to store data and higher current consumption to keep the device active.
    For AoA, we took 512 I/Q samples in each AoA packet, and 16 samples in 4us, that means 128us to finish sampling . On top of that we have application overhead, so it would be probably 2~3ms for a AoA packet to be received, sampled, and calculated.

    We have a chapter in our user's guide regarding Localization:
    dev.ti.com/.../index.html

    For more information regarding AoA. Please take a look at the readme file for the example(aoa_reciever).

    Regarding the accuracy, I will have to get back to you later.
  • 0
    TI__Guru 56855 points
    The AoA and ToF application examples are provided as reference designs only, and are not characterized for accuracy. The accuracy and latency in a productized application will be very much up to the system developers as they navigate the tradeoffs between system complexity, size, cost, and energy usage. However, we can say the following about accuracy:

    AoA:
    Angle measurement accuracy when averaged over 3 packets at 3 different frequencies using the AoA boosterpack is +/- 4 degrees
    Inter element distance, frequency diversity, and averaging over packets will impact performance (latency and accuracy)

    ToF:
    The current solution has a resolution of 18.75m and a variance that is also around 18.75m. Taking the average value of many measurements can improve the accuracy to an almost arbitrarily good level at the expense of latency and energy. As an example, if you average 500 measurements (about 250ms), you will have a 95% confidence interval of 4m. If you have calibrated your setup correctly, this interval will be centered on the correct distance, giving you an accuracy of +/-2m with 95% confidence level. We expect to improve the variance in future SW upgrades, allowing faster and/or more accurate results.

    Common for both AoA and ToF is that multipath transmission and reflections are the main challenges that limit accuracy in most real environments. The numbers given above assume a “clean” environment where only the intrinsic measurement uncertainty of our AoA and ToF systems contribute to the error.