• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 3 answers
  • Subscribers 31 subscribers
  • Views 724 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.

Elevation estimation using Capon beamforming

Resmi Johnson
Prodigy 75 points
Other Parts Discussed in Thread: IWR6843ISK, AWR6843AOP

Hello 

I am trying to understand the signal processing toolchain from the in the document '3D_people_counting_demo_implementation_guide'

path C:\ti\mmwave_industrial_toolbox_4_11_0\mmwave_industrial_toolbox_4_11_0\labs\People_Counting\docs

could you explain the relevance of the m_ind and n_ind in steering vector calculation ?

Also why phase rotation is applied only to azimuth and not to elevation.

  • 0
    TI__Mastermind 37780 points

    Hello,

    m_ind and n_ind are used to input the virtual antenna locations in the antenna array. Looking at the antenna pattern of the IWR6843ISK and/or IWR6843ODS will show this. See figures 17-18 in the document referenced.

    The steering vectors are calculated from the expected radar return for reflections at a particular angle. Since the signal received at each channel will depend on the virtual antenna locations, we need to know the positions of each virtual antenna channel to calculate the steering vector.

    The phase rotation happens because on the ODS platform, some receivers are fed from a different direction than others. See the text below figure 17 for more detail.

    "The virtual antenna coordinates and the order of antennas in Radar Cube memory generated by the Range DPU are shown in Figure 18. Numbers in the circles represent the antenna order. Note that this order holds if the transmit antenna order in the MIMO scheme matches their physical order. On the ODS EVM the Rx antennas Rx1 and, Rx4 are fed from the opposite side compared to Rx2 and Rx3, therefore the phase rotation of 180 degrees has to be applied either to Rx1 and Rx4 or to Rx2 and Rx3."

    Best,

    Nate

  • 0
    TI__Intellectual 790 points

    Hello,

    All the generated steering vectors (pointing to any arbitrary azimuth or elevation angle) should have antenna phase rotation and board-related phase bias (i.e., calibration coefficients). At the implementation level:

    • The elevation steering vectors are calculated by multiplying pre-calculated elevation steering vectors with the azimuth steering vector corresponding to the detected point.

    Hence, in order to incorporate the phase rotation and board-related phase bias only once, we should add these coefficients in one of the pre-calculated steering vectors (azimuth/elevation). Since the azimuth steering vector is used in the previous stage (when generating the range-azimuth heatmap), the coefficients are only included in the azimuth steering vector. Hence, when multiplying the pre-calculated vectors to point to an arbitrary angle, the overall steering vector will include the coefficients only once.

    Regards.

    Muhammet

  • 0 in reply to Nathan Block
    Prodigy 75 points

    Thank Nathan for the reply. I understood the relevance of m_ind and n_ind in steering vector calculation

  • 0 in reply to Muhammet Yanik
    Prodigy 75 points

    Hello

    Thanks for the explanation that you gave wrt phase rotation.

    With respect to elevation estimation , could you please confirm my understanding as :

    For each detection point in range azimuth plane using CFAR(from step 1 processing), we will be creating a 1D spectrum of  elevation by 1. create elevation steering vector for all mu (i from o to NE-1) using equation

    2. create azimuth steering vector for nu corresponds to azimuth point from CFAR using equation

    3. multiply the above two steering vector to get elevation steering vector

    4.elevation spectrum is calculated by multiplying the elevation steering vector (from step3) and spatial co variance matrix and will be further used to find the elevation angles.

    I am using AWR6843AOP .Please confirm whether my understanding on elevation estimation is correct or else please provide more info

    Also could you provide more info on the finding the spatial covariance matrix (12*12) for elevation estimation  .It was not explained in detail in '3D People Counting Demo Software Implementation Guide' (in the implementation level).

    Thanks and Regards

    Resmi Johnson

  • 0 in reply to Resmi Johnson
    TI__Intellectual 790 points

    Hello Resmi.

    Yes, all the steps you listed above are correct. You can see the details of covariance estimation at the following section of the '3D People Counting Demo Software Implementation Guide'. The document version I am referring is at the toolbox 4.11.0.

    - 6.1.2.1.2 Spatial Covariance Matrix Estimation and Inversion

    You can also refer to the following reference design (Section 2.3.2 Low-Level Processing) for more information:

    - https://www.ti.com/lit/ug/tidue71d/tidue71d.pdf?ts=1661359988710&ref_url=https%253A%252F%252Fwww.google.com%252F

    Regards.

    Muhammet

  • 0 in reply to Muhammet Yanik
    Prodigy 75 points

    Thank you for the detailed explanation.

    Regards

    Resmi Johnson