2D Loop Coils based Magnetic-field FoCusing (MFC) by Machine Learning

A machine learning-based Magnetic-field FoCusing (MFC) that simulates and learns A-Matrix, corresponding to B = AI, thereby enabling magnetic field focusing on a specific point is proposed in this paper. The conventional method of directly calculating the complex A-Matrix to solve problems has the disadvantage of rapidly increasing computation time due to the exponential increase in matrix dimensions and variables with the addition of magnetic field focusing points, increase in coil numbers, and expansion from two-dimension (2D) to three-dimension (3D). To address this, MFC method by training a A-matrix by using machine learning to effectively construct the A-Matrix under given 2D Planar Coil array conditions is utilized instead of conventional formulaic analysis based on Biot-Savart’s law. The performance of the proposed machine learning based MFC was verified through simulations that demonstrated magnetic field focusing in 2D 3x3 and 4x4 coil arrays.