• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 30 subscribers
  • Views 1161 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.

AWRL1432BOOST-BSD: DopplerDPU log2 Magnitude's dstAcnt

Hsu Yuri
Prodigy 60 points
Part Number: AWRL1432BOOST-BSD

Tool/software:

Dear TI Team,

I would like to inquire about the last stage in the Doppler DPU, specifically the log2 magnitude computation.

We can observe the following configuration:

hwaParamCfg[paramsetIdx].source.srcAcnt = 2 * (cfg->staticCfg.numDopplerChirps) - 1;

This is because there are two range bins, so:

  • R0: DopplerBin 0 ~ 127
  • R1: DopplerBin 0 ~ 127

The goal is to apply log2 magnitude to all Doppler bins across these two ranges, ensuring that the detection matrix is in log scale.

However, I noticed that dstAcnt is configured as:

hwaParamCfg[paramsetIdx].dest.dstAcnt = 2 * (cfg->staticCfg.numVirtualAntennas * cfg->staticCfg.numDopplerChirps) - 1;

I would like to ask why dstAcnt is additionally multiplied by cfg->staticCfg.numVirtualAntennas?

Thank you for your clarification.

Best regards,
Yuri

  • 0
    Prodigy 60 points

    Also 
    i want to ask about the reason of 
    Doppler FFT -> ABS Sum -> Log2 Magnitude -> Detection matrix 

    Why not:
    Doppler FFT -> Log2 Magnitude -> Sum -> Detection matrix 

    Thanks

    Yuri

  • 0 in reply to Hsu Yuri
    Prodigy 60 points

    Since the Sum Abs in this case is configured as

    hwaParamCfg[paramsetIdx].accelModeArgs.fftMode.magLogEn = HWA_FFT_MODE_MAGNITUDE_ONLY_ENABLED; // Enable magnitude operation only hwaParamCfg[paramsetIdx].accelModeArgs.fftMode.fftOutMode = HWA_FFT_MODE_OUTPUT_SUM_STATS; // Store sum statistics output in HWA memory

    This means that the I/Q signals from different antennas are directly summed together.

    In the next stage—log2 magnitude processing:

    hwaParamCfg[paramsetIdx].accelModeArgs.fftMode.magLogEn = HWA_FFT_MODE_MAGNITUDE_LOG2_ENABLED; hwaParamCfg[paramsetIdx].accelModeArgs.fftMode.fftOutMode = HWA_FFT_MODE_OUTPUT_DEFAULT;

    At this point, the I/Q signal values output from the Sum stage are processed to obtain Magnitude (Abs) and Log2.

    But does this mean that the detection matrix for beamforming at 0 degrees is essentially being formed?
    Since the I/Q signals from all antennas are directly summed here, isn't this equivalent to performing beamforming at 0 degrees?

    Additionally, why is it not done by first taking the absolute value (Abs) and summing the results of the Doppler FFT to obtain the Magnitude, and then applying Log2?

    Yuri
  • 0 in reply to Hsu Yuri
    TI__Expert 4250 points

    Hello Yuri, 

    Let me look into this and get back to you 

  • +1 in reply to Hsu Yuri
    TI__Expert 4250 points

    Please find answer to this question below 

    I want to ask about the reason of 
    Doppler FFT -> ABS Sum -> Log2 Magnitude -> Detection matrix 

    Why not:
    Doppler FFT -> Log2 Magnitude -> Sum -> Detection matrix

    If we do log2 first before the sum we loose precision and information about the values. To get a more accurate answer we do sum and then log2

  • 0 in reply to Zorah Ahmad
    Prodigy 60 points

    Dear Zorah
    There still has a problem about why dstAcnt is additionally multiplied by cfg->staticCfg.numVirtualAntennas?


    Thanks for your patience.

  • +1 in reply to Hsu Yuri
    TI__Expert 4250 points

    Hello Yuri, 

    A single Paramset is used to apply log2 to the two Doppler FFTs across all the three Rx antennas hence dstAcnt is additionally multiplied by cfg->staticCfg.numVirtualAntennas? .