Open AccessHybrid Tikhonov source current reconstruction method for large‐scale problems
Author(s) -
Bod Mohammad,
Sarraf Reza,
Moradi Gholamreza,
Jafargholi Amir,
Moallemizadeh Amir
Publication year - 2018
Publication title -
iet microwaves, antennas and propagation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.555
H-Index - 69
eISSN - 1751-8733
pISSN - 1751-8725
DOI - 10.1049/iet-map.2017.0264
Subject(s) - tikhonov regularization , krylov subspace , solver , transformation (genetics) , mathematics , algorithm , iterative method , mathematical optimization , computer science , mathematical analysis , inverse problem , biochemistry , chemistry , gene
Source current reconstruction method (SRM) is a field transformation technique in which the equivalent currents at the antenna aperture are reconstructed from the known radiated fields. To stabilise SRM equations and reduce the effect of measurement errors and noise, the Tikhonov SRM has been frequently used. However, finding the regularisation parameter and solving the Tikhonov SRM equation have a high computational cost especially for large‐scale problems or in the presence of both electric and magnetic currents. Here, a hybrid regularised least‐square minimum residual (LSMR) technique is proposed to implement the Tikhonov SRM equation. In this method, the Tikhonov SRM is applied on the Krylov subspace that is created by the LSMR iterative solver. The regularisation parameter of the hybrid Tikhonov equation is determined at each iteration of the projected problem with the generalised cross‐validation method. Results of current reconstruction and near‐field/far‐field transformation with the proposed hybrid Tikhonov SRM are reported in this study and compared with the standard Tikhonov SRM. It is shown in both simulation and measurement examples that a significant improvement in the computational cost is realised with the proposed hybrid method while all of the benefits of the traditional method are preserved.