An Efficient Data Reconstruction Method for Broadband Planar Near-Field Measurements Based on the Field Distribution Similarity

Abstract

This article presents an effective E-field reconstruction method for the near-field measurement system. The spatial convolution and field similarity theory are used to form two observation planes with similar E-field distributions. Then the clustering analysis and Voronoi cell classification are introduced to interpolate the defective E-field dataset at the target frequency using the similar dataset at another frequency and observation plane, which improves the data utilization efficiency in the frequency domain. The key purpose of the proposed method is to effectively reconstruct the defective planar E-field influenced by the under-sampling in the multi-frequency test, saving the retesting time cost of the E-field at the target frequency point and improving the utilization efficiency of the multi-frequency data. From the simulation and measurement results, it can be known that the reconstructed correlation coefficients are greater than 0.98, implying the high accuracy and good effectiveness of the proposed method. More importantly, the measurement time cost of the proposed method, which is 70 minutes for the measured results, can be half of that of the two-time measurement under the same frequency and observation distance, greatly saving the time cost for the electrically large antenna.