Achievable sidelobe level of the radiation pattern of a linear antenna array thinned by the genetic algorithm

Thinning of antenna arrays has been a popular topic for the last several decades. This paper suggests a genetic algorithm as an instrument for antenna array thinning. The algorithm with a deliberately chosen fitness function allows synthesizing thinned linear antenna arrays with low peak sidelobe level (SLL) while maintaining the half-power beamwidth (HPBW) of a full linear antenna array. A fitness function is proposed that allows determining by the genetic algorithm the location of the elements of a thinned antenna array, providing for a given amplitude distribution and number of elements a minimum SLL with a slight expansion of the main lobe of the radiation pattern. Based on results from existing papers in the field and known approaches to antenna array thinning, a classification of thinning types is introduced. The optimal thinning type for a linear thinned antenna array is determined on the basis of a maximum attainable SLL. It is shown that edge thinning of antenna arrays allows for obtaining a lower SLL than stochastic thinning. With the optimal thinning coefficient, the SLL of the radiation pattern of the full antenna array is higher than that of the thinned one with almost the same width of the main lobe. For the first time, the effect of the thinning coefficient on main directional pattern characteristics, such as peak SLL and HPBW, is discussed for a number of amplitude distributions.