PPT Slide
Controlling the Adapted Pattern Sidelobes
The adaptive algorithm makes no attempt to cancel Ideal thermal noise since it is not correlated between channels. The easiest way to control sidelobes is to add Artificial Ideal Thermal Noise.
Since the Covariance Matrix for Ideal Thermal Noise is a diagonal matrix, we can add to the sample covariance matrix a diagonal matrix which dominates the nuisance noise correlation.
This “Diagonal Loading“ Matrix makes the noise correlations from the finite sample appear to be weaker. This lessens there effect on the adapted pattern.