Room Temperature Magnetic Field Learning with Optically Readout Single NV-Centers

We propose and test a Magnetic Field Learning (MFL) protocol for high-resolution, high dynamic range and high-sensitivity magnetometry with a single NV-center electron spin. Our approach leverages recent proposals that analyze the benefits of adopting classical machine learning to post-process quantum data in quantum sensing protocols [1]. MFL was tested at room-temperature using a setup detecting state-dependent fluorescence via confocal microscopy, and a microwave controlled NV defect in bulk diamond as a sensor [2]. This setup senses the intensity of a magnetic field B in the proximity of the quantum sensor via Ramsey interferometry [3]. In our protocol, the Ramsey precession time τ is adaptively chosen at each step via an efficient particle guess heuristic [1].