A Numerical Investigation of the Application of Compressed Sensing in Spherical Mode Filtering for Near-or Far-Field Antenna Measurements

This paper presents the results of an extensive numerical investigation into the spherical mode filtering method and the application of Compressed Sensing (CS) for near- or farfield two-dimensional antenna pattern measurements. This study is an extension of the authors' prior work [1], [2] that expands the findings to the generalized case. When measuring an antenna pattern at a near-, intermediate, or true far-field distance, spherical mode filtering of an intentional offset antenna very effectively removes parasitically coupled multipath reflections from the test environment. The CS algorithm further enhances this technique by allowing the data to be sampled on an irregularly spaced spherical grid comprising only a few percent of the measurement points required by a conventional equiangular, Nyquist sampled, spherical acquisition. This investigation uses full-wave three-dimensional computational electromagnetic simulated data to examine the spherical mode filtering technique, addressing questions about the mechanism of modal separation, i.e. the mode orthogonalization, and sparsification facilitated by the translation of origins of the measured antenna pattern to the measurement rotation center. Additionally, the paper discusses potential limitations of the approach, identifying scenarios where modal filtering may be less effective-issues previously noted for one-dimensional cylindrical mode filtering but not yet documented in existing open literature for the analogous 2-D spherical case.