Sidelobe optimization of Skobelev Networks

A new optimization method for computing the coupling coefficients of the directional couplers constituting a Skobelev Network is presented in this paper. Skobelev Networks (also known as chessboard networks), typically used in Limited Field Of View (LFOV) applications, present a modular structure composed by N stages feeding linear arrays partitioned into several overlapped sub-arrays. Each network stage is made up of directional couplers whose coupling coefficients are determined through a constrained optimization routine based on deformed polyhedron methods. In this work an alternative optimization is proposed. The new procedure consists in two steps: 1) computation of the Skobelev Network Scattering Matrix, 2) constrained optimization of directional coupler coefficients adopting a Sequential Quadratic Programming (SQP) method. The results obtained employing this new routine show a remarkable reduction in the sub-array radiation pattern sidelobe level (from 4.7 dB for N = 1 up to 6.49 dB for N = 4) with negligible effects on the main beam.