FIG. 3 shows the simulated optimized reactive match data curve set of the six antenna elements of the sub-array 110 (FIG. 4) over a frequency range of 2.1-2.7 GHz by ANSYS HFSS simulation. Curve 301 corresponds to element-1 in the sub-array 110 of FIG. 4, curve 302 corresponds to element-2 in the sub-array 110 of FIG. 4, curve 303 corresponds to element-3 in the sub-array 110 of FIG. 4, curve 304 corresponds to element-4 in the sub-array 110 of FIG. 4, curve 305 corresponds to element-5 in the sub-array 110 of FIG. 4, and curve 306 corresponds to element-6 in the sub-array 110 of FIG. 4. The reactive match is defined as the return loss of the individual element when all the elements are excited simultaneously (i.e. a real operation condition). In other words, the reactive return signal is a vectorial sum of the self-return signal of the antenna element plus the incoming signals attributed to mutual coupling from all the surrounding antenna elements. Thus, a good match of any given antenna element in an array environment means that the self-match signal of that element is cancelled out by the total signal of all the mutual coupling. Under the circumstance of roughly identical radiator design across the entire aperture, the self-match of all the individual elements is about the same, while the total signal attributed to mutual coupling for each element is be different from element to element due to the non-symmetrical geometrical positions of the elements on the aperture. This may manifest as relatively poor reactive matching at the corner elements 200a, as shown by curve 306 in FIG. 3, and as discussed further below.