Computationally-Effective Optimal Excitation Matching for the Synthesis of Large Monopulse Arrays

Antenna arrays able to generate two different patterns are widely used in tracking radar systems [1]. Optimal (in the Dolph-Chebyshev sense) sum [2] and difference patterns [3] can be generated by using two independent feed networks. Unfortunately, such a situation generally turns out to be impracticable because of its costs, the occupied physical space, the circuit complexity, and electromagnetic interferences. Thus, starting from the optimal sum pattern a sub-optimal solution for the difference pattern is usually synthesized by means of the sub-array technique. The array elements are grouped in sub-arrays properly weighted for matching the constrains of the difference beam. Finding the best elements grouping and the sub-array weights is a complex and challenging research topic, especially when dealing with large arrays. As far as linear arrays are concerned, McNamara proposed in [4] an analytical method for determining the �gbest compromise�h difference pattern. Unfortunately, when the ratio between the elements of the array and sub-arrays increases, such a technique exhibits several limitations mainly due to the ill-conditioning of the problem and the computational costs due to exhaustive evaluations. A non-negligible saving might be achieved by applying optimization algorithms (see for instance [5] and [6]) aimed at minimizing a suitable cost function. Notwithstanding, optimization-based approaches still appear computationally expensive when dealing with large arrays because of wide dimension of solution space to be sampled. In order to properly deal with these computational issues, this contribution presents an innovative approach based on an optimal excitation matching procedure. By exploiting the relationship between independently-optimal sum and difference patterns, the dimension of the solution space is considerably reduced and efficiently sampled by taking into account the presence of array elements more suitable to change sub-array membership. In the following, the proposed technique is described pointing out, through a representative case, its potentialities and effectiveness in dealing with large arrays. This is the author's version of the final version available at IEEE.