Analysis by Systemic Approach of Magnetic Dipole Source Location Performances by Using an IoT Software Gradiometer Head

Abstract

Based on the development of numerous commercial cheap or low noise magnetometers, some publications feature systems of magnetic anomaly detection. They are meanly based on matrix inversion gradients and are built with multi-magnetometer single head in order to derive spatially the sensed magnetic field. Meanwhile, these examples often suffer from a lack of true performance analysis, based on a &#x201C;systemic approach.&#x201D; As an example, they take account partially of the main sensor, mechanical or conditioning electronic characteristics. In that way, we propose a theoretical modeling of 1D and 2D magnetic head detection capabilities that could be extended to 3D. We show that the distance of detection could be quantified with the help of three main parameters: the baseline, the sensor mismatch in terms of sensitivity (V/T) or misalignment, and a defined signal-to-noise ratio of the measurement (<inline-formula> <tex-math notation="LaTeX">$\textit {SNR}_{m}$ </tex-math></inline-formula>). With a reasonable built system (sensitivity disparity of 0.6 &#x0025;, misalignment lower than some (&#x00B0;) and <inline-formula> <tex-math notation="LaTeX">$\textit {SNR}_{m}$ </tex-math></inline-formula> of 120 dB), a magnetic dipole could be localized and quantified, from <inline-formula> <tex-math notation="LaTeX">$\times1$ </tex-math></inline-formula> up to <inline-formula> <tex-math notation="LaTeX">$\times20$ </tex-math></inline-formula> gradiometer base line with a relative error lower than 30 &#x0025;, and from <inline-formula> <tex-math notation="LaTeX">$\times3$ </tex-math></inline-formula> up to <inline-formula> <tex-math notation="LaTeX">$\times50$ </tex-math></inline-formula> with a <inline-formula> <tex-math notation="LaTeX">$\textit {SNR}_{m}$ </tex-math></inline-formula> of 160 dB, as an example. Notice that the highest distance of detection is mainly bounded by the <inline-formula> <tex-math notation="LaTeX">$\textit {SNR}_{m}$ </tex-math></inline-formula> value and the intrinsic head spatial transfer function. At a close distance, we have enlarged significantly the surface of the source localization, by using optimization algorithms when classical gradiometric method appears non-functional. To exemplify this systemic approach, we have built and characterized a setup. It is based on the IoT principles, where acquired data are pushed over-the-air through a gateway. Finally, they are post-processed in quasi-real time with a laptop.