Application of Compressed Sensing to Antenna Far-field Calibration in an Extrapolation Range

Antenna gain calibration in an extrapolation range is the prevailing method used by most metrology labs and is considered the most accurate method for acquiring far-field gain in an indoor chamber. This method requires one of the antennas to move along a linear track to collect the responses. The responses as a function of distance exhibit ripples due to multipath reflections between the two antenna and in the measurement chamber. Post processing is needed to remove these ripples before fitting to a generalized equation to extrapolate the responses to the far field. The ripples are traditionally smoothed out by using a moving average method, and this requires collecting data densely to sufficiently average out the ripples. Recent proposals using k-space filtering can relax the sampling requirement to a certain extent. However, when interactions between antennas manifest at high spectral frequencies, dense sampling becomes essential to prevent aliasing in the spectrum domain. In this study, we propose to use Compressed Sensing (CS) to recover the spectrum coefficients, and then apply filtering in the k-space. CS scheme can significantly reduce the number of sampling points required and speed up the measurements while maintaining the accuracy of the extrapolation method.