Fast backward singular value decomposition (SVD) algorithm for magnetocardiographic signal reconstruction from pulsed atomic magnetometer data

In a pulse pump Rb atomic magnetometer, the magnetic field is associated with the Larmor frequency of the free induction decay (FID) signal. The reconstruction of the magnetic field from the collected signal, thereby, is crucial for magnetocardiography. In this study, we propose a backward singular value decomposition (BSVD) method for fast reconstruction of a magnetocardiographic signal. Experiments on the simulated and real data were performed to estimate its potential advantages over previous approaches, such as the fast Fourier transform (FFT) method, the zero-crossing means (ZM) method, etc. The results show the high accuracy of the BSVD method compared with other methods. More importantly, the BSVD method requires less sampled data than other methods while ensuring the accuracy. With the help of it, the recording time can be greatly reduced from the initial 3.6m s to the present 0.6m s. Thus, the time resolution of the magnetocardiograph could reach 2m s which is equivalent to that of conventional electrocardiogragh. This will bring the atomic magnetocardiography more practicable in clinic application.