Nonlinear least-square solution to flat-top pattern synthesis using arbitrary linear array

This paper presents a new approach for synthesizing flat-top patterns, based on the least-square error criterion. The cost function is formulated according to the amplitude approximation error without phase constraint. The optimal array weight is obtained by using the Levenberg-Marquardt nonlinear optimization algorithm. Simulations are performed to compare the proposed approach with the Woodward-Lawson method and two recent methods using minimax and adaptive array techniques, respectively. The results indicate that the approach is effective in sidelobe control and synthesizing prespecified patterns for arbitrary linear arrays.