Hi,
We observe strange doppler ghosts when measuring. The problem gets bigger with stronger or longer targets and seems to repeat it self in Doppler direction at fixed intervals both above and below the target speed. The ghosts can be up to -7 dB with respect to the target itself in strength. Multi paths would show up in range domain, be they are not visible. Is this a problem that sounds familiar?
My prime suspect is jitter on the chirp timing with respect to ADC sampling moment. Looking at the timing settings, the idle time, ramp time and ADC start time can be set with 10 ns resolution. The ADC sampling clock can be configured to an arbitrary frequency. It is very challenging or impossible to realize this while respecting a 100% deterministic chirp timing and timing between the ramp start moment and sampling moment of the ADC. In the mmwave sensing estimator does not seem to make sure that jitter can occur as far as I can see.
I created a test in Matlab that samples a sine wave with a deterministic jitter timing interval [0.1, 0.1, 0.1, 0.11, 0.1, 0.1, 0.1, 0.11, 0.1, 0.1, 0.1, 0.11] and feed the samples through an FFT. This results exactly the same pattern, where the number of peaks is dependent on the periodicity of the jitter, but independent on the sine frequency.
Is there detailed information on the clocking of the ramp timing generator and the ADC sampling clock generation that could help in deriving a 100% deterministic timing or to quantify the error so we can maybe correct for it in the Fourier process?
The most detailed info I found was in the Programming Chirp Parameters in TI Radar Devices.
Is there any way we can determine if this is problem or rule it out? Any help is very much appreciated.
