Near-Field Pattern Synthesis for Sparse Focusing Antenna Arrays Based on Bayesian Compressive Sensing and Convex Optimization

An effective method based on Bayesian compressive sensing (BCS) and convex optimization for near-field sparse array synthesis is presented in this paper. An algorithm to generate reference-shaped beams in the near-field region with controllable sidelobe levels is first proposed. Then, the multitask BC is modified and generalized to synthesize a near-field sparse array radiating a desired near-field pattern with the co-polarization component. After that, a postprocessing of the final array excitation is employed to put constraints on the minimum element spacing to make the sparse layout practicable. The degradation of the near-field pattern is mitigated through reestimating the array excitation by a convex optimization. Three numerical examples show the effectiveness of the proposed method with more than 50% of elements saved compared to the uniformly distributed layout. The comparison to the result obtained by a full-wave simulator FEKO is also presented to demonstrate the validity of this method considering strong antenna mutual couplings.