Speaker
Description
Quantum sensors exploit quantum properties of matter to measure and detect with high accuracy, sensitivity and/or resolution. Among the different types of sensors available, sensors based on NV centers have the additional advantage that they can be used at room temperature. This characteristic opens the door to develop new measuring, inspection and diagnostic equipment for processes and products from living organisms to production lines. The difficulty relies on the complexity of the obtained signal. The signals generated by NV center-based sensors are characterised by being very weak, i.e. of low intensity, and at the same time have a strong noise component. In this work we present a mixed approach linking a theoretical modelling and selection of the excitation transfer sequences and the use of advanced processing methods. The artificial intelligence processing system exploits the local relation codified via convolutional neural networks and the time features via recurrent neural networks, specifically long short-term memory networks.
The hyperfine coupling parameters are predicted with an error of less than 5 KHz for the 96,78% of test samples.
