Reconstruction of Faulty-Free Antenna Pattern in the Presence of Fault Elements of a Transmitarray in Planar-Near-Field-Measurement

This paper presents an effective source reconstruction method for planar near-field-measurement-based diagnosis. The proposed method uses spatial convolution to extrapolate the two planar far fields under the two frequency points, where one is the over-sampled low-frequency far field, the other is the under-sampled high-frequency far field. Based on the progressive similarity theory, the two far fields are highly similar and the under-sampled one at high frequency can be assimilated by the over-sampled one at low frequency. In this regard, the Kmeans clustering and Voronoi diagram are used to realize the data assimilation, and the reconstructed far field is then back projected onto the extreme near-field plane by the deconvolution method. Finally, the far-field pattern can be obtained by the near-to-far-field transformation algorithm, having good accuracy compared with the simulated one. More importantly, the proposed method can save the acquisition time significantly and improve the efficiency of the diagnosis measurement system.