CFARDectector2D Error using fzero: Initial function value must be finite and real.

Hi,

I am trying to create a CFAR range Doppler map for FMCW radar.

Offending code:

cutidx = [];

for m = 1:n_sweeps

for n = 1:n_samples

cutidx = [cutidx,[n;m]];

end

end

ncutcells = size(cutidx,2)

[up_detections, upth] = CFAR(abs(fftshift(IQ2D)), cutidx);

My 2D FFT result is IQ2D, dimensions 200 samples x 512 sweeps. I set up the cutidx as shown in the example. Transposing the input has no effect on the error:

Error using fzero

Initial function value must be finite and real.

Error in phased.internal.AbstractDetector/calcWGNThresholdFactor (line 353)

alpha = abs(fzero(@(x)SOCAWGNThresholdFactor(abs(x)),cast(CAThreshold,'double')));

Error in phased.internal.AbstractDetector/setupImpl (line 216)

obj.pFactor = calcWGNThresholdFactor(obj);

Error in phased.CFARDetector2D/setupImpl (line 243)

[e-mail](obj);

Error in cfar_rd_sawtooth (line 80)

[up_detections, upth] = CFAR(abs(fftshift(IQ2D)), cutidx);

Please find script and data attached.

Answer 1

[Dayalan Nair]

The issue was that the cells under test must be training + guard cells away from the edges of the data set