An Adaptive Data Acquisition and Clustering Technique to Enhance the Speed of Spherical Near-Field Antenna Measurements

Abstract

This letter presents a new approach for adaptive spherical near-field (NF) antenna measurement. The proposed method begins with a number of initial points and sequentially focuses on the areas with a highly dynamic NF pattern. Thereafter, the source reconstruction method is utilized to calculate the equivalent magnetic and electric currents on the surface of an ellipsoid that encompasses an antenna under test. The equivalent sources are used to compute the far-field pattern of the antenna. The comparison of the adaptive algorithm with the uniform sampling indicates that the number of the required samples is decreased significantly using the adaptive method. The adaptive data acquisition can also be used in case of uniform sampling to remove the redundant samples and accelerate the source reconstruction method. Since the newly added point is not necessarily laid on the measurement points, the cubic spline interpolation technique is employed to compute the value of the field. Besides, a machine learning algorithm based on k-means clustering is applied to the uniformly sampled data to determine different clusters of data. Thus, for every new point, the cluster to which the data point belongs can be determined, and only the values of that cluster are used to calculate the value of the new point.