Array Thinning Through Binary Sequences

Thinning is a standard technique to reduce the power consumption, hardware complexity, cost, and weight of large arrays [1][2]. However, array thinning does not allow a suitable control of the peak sidelobe level (PSL), as for filled arrays, with a reduction of the array performances [1][2]. In the literature, several techniques have been proposed for the design of thinned arrays with low sidelobes [1-][4]. In a chronological order, arrays were at first thinned by deterministic techniques, but no significant advantages on the PSL reduction have been obtained over random approaches [1]. Successively, stochastic optimization techniques have been widely and successfully applied [3][4]. However, such techniques present high computational costs when applied to large arrays, and they do not allow a-priori estimates of the expected performances for a given aperture size and thinning factor [2]. More recently, thinned arrays have been synthesized by exploiting the autocorrelation properties of binary sequences derived from Difference Sets (DSs) [2][5]. Such an approach has several advantages over traditional thinning techniques. It is computationally efficient and its performances are predictable [2]. However, DSs exist only for a limited subset of configurations, and their features cannot be exploited in a general way [2][5]. In order to apply the same strategy to a wider set of array configurations, sequences which are sub-optimal with respect to DSs can be considered. For instance, Almost Difference Sets (ADSs) [6] are good candidates for large array thinning, since they represent a generalization of DSs with several similar properties [6]. This paper is aimed at analyzing the applicability of ADSs to the design of thinned linear arrays, as well as the effectiveness of such an array synthesis technique in terms of beam-pattern features. Selected numerical results are provided to assess the performances of the proposed approach also in comparison with state-of-the-art thinning techniques with predictable PSL behaviour.