IWR1443: rangeDopplerLogMagMatrix data comparison with fftOut1D data

Hi Nitin,

Thank you for your reply. I am familiar with the high level discussion presented in

C:\ti\mmwave_sdk_02_01_00_04\packages\ti\demo\xwr14xx\mmw\docs\doxygen\html\index.html

I guess I should define what I mean my comparison.  I am aware that the 1d-fft output is 16 bit complex and the range doppler log magnitude is 16 bit real.  This just means the 1d-fft output just has to have is complex absolute value generated for comparison purposes and a log2 then as well.  But how?

I had done a little experiment last week with the sensor configured with just one antenna and four chirps per frame.  I had a scatterer move in front of the sensor and I collected 25 frames. I then selected the peak magnitudes for the four chirps per frame from the 1d-ouput.  I had also selected the peak maglog values in the rangedopplerlogmagnitude array.  These values are shown in figure comp.gif.  The red curve is the rangedopplerlogmagnitude data and the other four curves are the 1d-fft data.  Figure rangeDoppler.gif contains just the rangedopplerlogmagnitude by itself.  Yes, the peak value for each point in the comp.gif plot will not be at the same range index value but close to it.  The include code illustrates how I arrived at the plotted data.

I just do not see how summing over the 1d-fft data using sum i = 0..N-1 {log2|x(i)|) will yield the data in the rangedopplerlogmagnitude array (in my example N=4).  I believe the values for the rangedopplerlogmagnitude array are generated in the *radarss.bin program which I do not have the source code for.

Can you please provide me with the part(s) I am missing?

Thanks.

Al

        MmwDemo_process2D(dataPathObj);
        /* 2nd Dimension FFT done! */
    mag0=0;
    mag1=0;
    mag2=0;
    mag3=0;
    mag4=0;    
    for(i = 0; i < dataPathObj->numRangeBins; i++)
        {
      a1=fftOut1D[i*dataPathObj->numDopplerBins].real;
      a2=fftOut1D[i*dataPathObj->numDopplerBins].imag;
      mag=a1*a1+a2*a2;
      if(mag>mag1)
        {
          mag1=mag;
        }
      a1=fftOut1D[i*dataPathObj->numDopplerBins+1].real;
      a2=fftOut1D[i*dataPathObj->numDopplerBins+1].imag;
      mag=a1*a1+a2*a2;
      if(mag>mag2)
        {
          mag2=mag;
        }          
      a1=fftOut1D[i*dataPathObj->numDopplerBins+2].real;
      a2=fftOut1D[i*dataPathObj->numDopplerBins+2].imag;
      mag=a1*a1+a2*a2;
      if(mag>mag3)
        {
          mag3=mag;
        }
      a1=fftOut1D[i*dataPathObj->numDopplerBins+3].real;
      a2=fftOut1D[i*dataPathObj->numDopplerBins+3].imag;
      mag=a1*a1+a2*a2;
      if(mag>mag4)
        {
          mag4=mag;
        }
      if(dataPathObj->rangeDopplerLogMagMatrix[i*dataPathObj->numDopplerBins]>mag0)
        {
          mag0=dataPathObj->rangeDopplerLogMagMatrix[i*dataPathObj->numDopplerBins];
          ind0=i;
        }
    }
    System_printf("%d %d %d %d %d\n",mag0,mag1,mag2,mag3,mag4);

Hi Allen,

@Allen: I believe the values for the rangedopplerlogmagnitude array are generated in the *radarss.bin program which I do not have the source code for.

The Radar Hardware Accelerator on board the IWR1443 device is used for the FFT and LogMagnitude operations and it is programmed in the MSS application code itself. Please look at the Radar Hardware Accelerator training video under mmwave-training-series for an introduction. For more details, refer to the Radar Hardware Accelerator user guide.

Regards

-Nitin