Other Parts Discussed in Thread: IWR6843AOP
Tool/software: Out of Box Demo
Asking a follow-up question as the original was never entirely answered.
The overarching question is to be informed on how the measurements output from the sensor are quantized, the answer to which I believe is rooted in the FFT size for position/velocity, and potentially the same for azimuth/elevation with some doubt.
I was at first under the impression that the range FFT size is the number of samples and the doppler FFT size is the number of chirps per frame, but analyzing some data returned by the sensor, this seem to perhaps not be the case. The TI sensing estimator on this chirp reports
- Maximum velocity 3.995
- Number of chirp loops 60
- velocity resolution 0.133
From this, if the doppler FFT has 60 bins I would expect the output to be quantized by 2 * 3.995 / 60 = 0.133, but in reality seems to be 0.124919, corresponding better to 64 FFT bins. Same comparison for range
- Maximum range 15.999
- Number of samples 255
- Range resolution 0.078
same calculation as before resulting in quantization interval of 15.999 / 255 = 0.0627 when data shows something more like 0.078. Furthermore, I was under the impressions that the angle-FFT is locked at 64 bins for each azimuth and elevation, given a reported max FOV +-60 (which should be +-90 in reality I think?) results in 1.875 degrees. Data shows differences as low as 0.01 for azimuth and 1.79 for elevation
Of course the calculation for resolution of range, doppler, angle is covered elsewhere, I have not found much regarding the implementation of the calculation. Can anyone shed some light on the details of the calculation, e.g. how FFT sizes are calculated and what steps are applied post FFT? If you can refer me to the correct place in the implementation, this is perfectly fine as well.
