| Title |
Artificial Neural Network-Based Magnetic-Field Focusing Using 2D Loop-Coil Arrays |
| Authors |
Min-Seung Kim ; Min-Seob Lee ; Young-Jin Park ; Eun-Soo Lee |
| DOI |
https://doi.org/10.6113/TKPE.2025.30.6.513 |
| ISSN |
1229-2214(pISSN), 2288-6281(eISSN) |
| Keywords |
Magnetic-field-FoCusing (MFC); Artificial Neural Network (ANN); Finite Element Method (FEM) |
| Abstract |
Conventional numerical analysis techniques based on the Biot?Savart law have been widely employed for magnetic field focusing (MFC) at specific target locations. However, these approaches suffer from exponentially increasing computational burden as the number of transmitting (Tx) coils and the dimensionality of the target space increase, which limits scalability and real-time applicability. In particular, the conventional requirement of employing more Tx coils than receiving (Rx) points makes practical MFC implementation challenging. To overcome these limitations, this paper proposes an artificial neural network (ANN)-based algorithm for efficient MFC using finite element method-generated magnetic field data. The proposed PyTorch-based ANN model is trained to predict magnetic field intensity and determine optimal current distributions for MFC operation. The effectiveness of the proposed approach was validated using a 4 × 4 Tx coil array and 9 × 9 target points, in which the number of Rx points exceeds the number of Tx coils. Results demonstrate that the proposed ANN-based MFC algorithm achieves accurate MFC, yielding over 99% prediction accuracy at designated target locations. |