Negative SNR in MATLAB sensorsig example

Hello everyone,

I have been using MATLABs sensorsig function. In one of their examples they stated that they wanted to use a SNR of 10dB. https://de.mathworks.com/help/phased/ref/sensorsig.html

Since that puzzled me quite a bit (usually a negative SNR is not very desirable as the noise power is overpowering the signal power) I generated a few matrices and tested those for my MVDR Beamformer and MUSIC algorithm.

For generating the matrices I created an array response of a 2-receiver radar using an operating frequency of 60GHz and a signal impinging on the ULA from -20° azimuth.

I used the following code:

fc = 60e9;
c = physconst('LightSpeed');
lambda = c/fc;

ula = phased.ULA('NumElements',2, 'ElementSpacing', lambda60/2);
pos = getElementPosition(ula)/lambda;

ang = [-20;0]
for SNR=-60:10:60
    [~,~, r]= sensorsig(pos, 50,ang, db2pow(SNR));
end

Interestingly, the algorithm performed excellent for negative SNR but very poorly for positive SNR.

As an example, please see my output for the MVDR Beamformer using -20dB for the SNR:

and now please compare it to the output of the same algorithm for a positive SNR of 20dB:

As you can clearly see, the algorithm detects the target perfectly at -20dB SNR (the peak of the PSD) and not at all for 20dB SNR.

Could someone please explain why this is the case?

Please correct me, if I am mistaken, but is this just some sort of a MATLAB conversion?

Thanking you kindly for your answers!

Best regards,

Nastassia