Model-Driven Compressive Spherical Array Sampling of Antenna Near Field Using Accelerated Pseudoskeleton Tensor Approximation

The antenna near-field samples (NFSs) are obtained by a circular array of probes (CAP) in spherical array sampling systems. In order to reduce sampling time, this article proposes a compressive spherical array sampling technique based on pseudoskeleton tensor approximation (PSTA). Specifically, Huygens’ principle and the method of moments (MoM) are applied to construct a tensor-based radiation model. Then, in order to realize the compression of NFSs and, thus, to reduce the sampling effort, the skeleton sampling positions of CAP are determined via PSTA of the tensor model, where NFSs in nonskeleton sampling positions can be computed from those in skeleton sampling positions. Moreover, a fast algorithm is developed to accelerate the PSTA of the tensor model. The ring antenna array and the patch antenna arrays are used to validate that the proposed method is applicable to antennas with different shapes or multibeam techniques. Numerical experiments show that the proposed method can determine the skeleton sampling positions with high computation efficiency and reduce the number of samples by 50% compared with uniform sampling.