Evaluating probabilistic and data-driven inference models for fiber-coupled NV-diamond temperature sensors

To meet such disparate measurement needs, a variety of temperature sensors have been developed. Although these devices vary greatly in their cost, size, weight and complexity, they almost all rely on well-established measurements of transport properties to infer temperature. Legacy technologies like platinum resistance thermometers and negative temperature coefficient thermistors have been relied upon for over a century to provide accurate and reproducible measurements over a broad range of temperature [2-4]. However, these sensors are prone to drift and require frequent re-calibrations to ensure high accuracy in critical use-cases resulting in increase cost of sensor ownership. In recent years, there has been a growing interest in developing alternative sensor technologies that can overcome the limitations of traditional technologies. The past decade has seen a burst of activity in nanophotonics [5], quantum optomechanics [6] and noise thermometry [7]. These technologies leverage telecomm industry's vast economies of scale along with precision measurement expertise developed for frequency metrology to enable fit-for-purpose, cost-effective measurement solutions. Development of an ultra-stable temperature sensor that shows minimal drift over decadal time spans or a field-deployable thermodynamic temperature sensor, likely based on quantum technologies could disrupt the calibration-centered metrology ecosystem of today [4, 5].