A genetic algorithm for optimizing the segmentation in subarrays of planar array antenna radars with adaptive digital beamforming

Modern phased array radars (PAR) should usually assure full performance in target detection and location also in presence of electronic counter measures (ECM). The radar signal processor of a PAR must provide then an adaptive technique to cancel the disturbance coming from jammers (A. Farina, 1992). The requirements on detection and location are particularly hard to fulfil when the power of jamming signal gets into the radar receiver from the main beam of the antenna pattern. To counteract the performance degradation due to main beam jammers (MBJ), the radar processor can adaptively combine digital signals coming from different subsets of the elementary radiators of the antenna, called subarrays, to form one or more beams with nulls in the directions of the jammers. This technique is known as adaptive digital beamforming (ADBF) (A. Farina, 1992). Fixed the complexity of the antenna (number of elementary radiators and subarrays), finding a set of subarrays that maximizes the detection and location performance of the PAR in presence of MBJs is a compelling challenge, because the function to optimise is irregular and includes a large number of parameters. Genetic algorithms are recognized as robust searching techniques: they are effective with very irregular functions and they guarantee much better efficiency than enumerative methods (D.E. Goldberg, 1989). They can explore different solutions in parallel, minimizing the probability to remain trapped in local maxima, and act directly on the functional to optimize, not on its derivates. Genetic algorithms have been used in various radar applications. In this paper an example of genetic algorithm for finding a optimized set of subarrays for ADBF is presented in details. Some results obtained with an antenna consisting in a square planar array of 64 radiators divided in 4 subarrays are also shown.