Hi,
What is the physical quantity and units of Range-Doppler Detection matrix given by OOB demo?
Is it power in mw, dB or magnitude in volt?
Thanks
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.
Hi,
What is the physical quantity and units of Range-Doppler Detection matrix given by OOB demo?
Is it power in mw, dB or magnitude in volt?
Thanks
Hi,
It is a relative quantity of power expressed in dB.
Best
Nate
Hi,
Thanks Nate..
Couple of follow up questions,
1.relative to what?
2. The value seems to be very high, is it log2 quantity in q9 format? Do I need to do a antilog (Q to decimal Value) to get the power in DB
Thanks
Hi Gayatri,
1.relative to what?
I believe it's estimated from the noise floor, but I can investigate this in greater detail.
2. The value seems to be very high, is it log2 quantity in q9 format? Do I need to do a antilog (Q to decimal Value) to get the power in DB
Are you referring to the values coming out of the radar device, or what's seen in the GUI demos online? There is some intermediate processing that happens between the two.
Best
Nate
Hi Nate,
Can you confirm what is the value relative to?
I am referring to the values coming out of the radar devices. These value seems to be very high, is it log2 quantity in q9 format? Do I need to do a antilog (Q to decimal Value) to get the power in DB? Can you please check?
Thanks
Hi Gayatri,
The values coming off the device are in a log format. To get them into a dB format in our visualizer tools, we run the following operations:
dB =
Note that I excluded the final term used by the source, which is present to make the visualizer work more uniformly in different scenarios. We have the 2^ceil(log2NumVirtAnt) / NumVirtAnt because we take an FFT across all the antennas to coherently combine the radar into a single measurement. As such, using 12 antennas requires us to run a 16 point FFT. Since the FFT happens in the HWA, it is scaled at each step by a factor of 2, which reduces power incorrectly by a factor of 1/4, because there are only 12 samples in the FFT (12/16 = 3/4. 1-3/4 =1/4)
The 20log10(2)/512 is to convert from Q9 to dB.
Note - E2E will be down from 9/28 to 10/2. Expect delays in responses accordingly.
Best,
Nate
Hi Nate,
Thanks for the reply. Can you please let me know what is the power relative to? Earlier you mentioned its estimated from the noise floor but in the post attached (last post in the thread) in the earlier comment mentioned that the power is relative to each other ( I didnt understand how that is measured).
Thanks
Hi Gayatri,
I stand corrected. The range-doppler matrix is not in fact a relative value. Rather, it's the direct output of 2D FFT + non-coherent summation across the antennas (which is a kind of absolute value of the magnitude range-Doppler spectrum). But the units are not linear (they're logarithmic), and the conversion to the dB scale is described above.
Best,
Nate