Radiated Near Field Prediction Based on Equivalent Source Reconstruction Method with Truncated Singular Value Decomposition

Chang Liu,Huapeng Zhao,Jun Hu

DOI: https://doi.org/10.1109/aps/ursi47566.2021.9704666

2021-01-01

Abstract:The equivalent source reconstruction method (ESRM) is a popular method to predict radiated near field. However, the original equivalent currents can't be reconstructed by solving the interaction matrix equation directly because the matrix equation is generally ill-conditioned. There will be a large error in the near field prediction, especially when the region of interest is very close to radiation source. To reduce the error of near field prediction, the truncated singular value decomposition is used to solve the interaction matrix equation. By ignoring small singular values of interaction matrix, its condition number is improved. Thus, the reconstructed equivalent currents are more similar to the original equivalent currents, and the error of prediction can be reduced significantly. Simulation results will be shown to illustrate the efficiency of the proposed method.