• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 894 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.

AWR1243BOOST: AoA estimation

Sanoal Machado
Prodigy 30 points
Part Number: AWR1243BOOST
Other Parts Discussed in Thread: AWR1243

Hi everyone,

We are currently working with the AWR1243+DEVPACK+TSW1400EVM. We are now analyzing the raw ADC data and reached to the point of angular estimation(Angle FFT).

In our implementation , we have defined the angle bins by dividing Max.FOV(90 deg)/No. of RX . Since we used only 1 TX and 4Rx for our test data,  we now have only 4 angle bins. Hence the step size between the angle bins is very large and we are not able to detect the precise angle for all the cases. Now we are also aware of the fact that, the angle resolution would have been better, if we had configured 2 TX which would have yielded in 2*4= 8 Rx virtual channels. However, we know that radars used in cars can have a small angular spacing/resolution (< 1 deg) even with a 1 TX setup. Is this solution achieved by chirp configuration or by some DSP method?

Thank you in advance for your help!

  • 0
    TI__Guru*** 148884 points
    Hi,

    Here is a useful thread that explains how angular resolution is computed.

    e2e.ti.com/.../681263

    angular resolution <1 deg can't be achieved with 1TX. One will need more channels.

    Do you have a source for the information "we know that radars used in cars can have a small angular spacing/resolution (< 1 deg) even with a 1 TX setup."?

    Thank you
    Cesar
  • 0
    Expert 1765 points
    Hi,
    You can try to use Super Resolution algorithm like MUSIC and such.
    But I believe even with these algorithms, it is not straight forward to get 1 deg resolution with 4 Virtual antennas and 90 deg Fov.

    The limitation of resolution mentioned by Cesar is valid only when you are performing AoA by FFT.

    Hope it helped.
    Santhana Raj
  • 0 in reply to Santhana Raj
    Prodigy 30 points

    Hi,

    So, just to have a complete understanding from your response:

    I have seen from the TI-training series that angular resolution can be defined as:

    Theta_res = (2/N) [rad]. 

    So, for example, if we use a set up of 1TX and 4RX antennas, angular resolution will be limited to:

    (1/2)[rad]  = 28.65 [deg] .

    If I have used an entirely FFT based algorithm (Angle-FFT), does this mean that the moment of defining the angular bins, I can only achieve N angular bins with Theta_res step size ?. (Right now, in the algorithm of angle FFT, I am calculating the angular spacing/step size = max FOV(90)/4) = 22.5 deg Which should be the correct way of calculating the angular spacing/step size?.

  • 0 in reply to Sanoal Machado
    Expert 1765 points
    Hi,

    Your understanding is correct regarding the formula when using FFT based AoA.

    The actual formula is:

    Theta_res = lambda/(N d cos(theta)),

    where, Lambda denotes the wavelength, and d is the spacing between antennas.
    For EVM, Lambda/d = 2; For this spacing the antennas FoV is 90 degrees. Hence the formula is reduced for FoV of 90 degrees.

    Your formula of "angular spacing/step size" is a approximation. Note that in the original formula there is a factor of theta in the formula. This means that the Resolution is not uniformly linear as theta varies. The resolution is max at 0 degrees. This maximum resolution at 0 degree is only calculated by the formula (2/N). Your formula assumes more or less an uniform resolution in the FoV. Still I beleive, this is a good approximation for fast calculation.

    Hope it helps.
    Santhana Raj

    PS: i would advise you to post new questions rather than posting here as the issue is being marked as resolved.