Supporting data for "Coupling In-Sensor and In-memory Computing with Diamond Nitrogen-Vacancy Sensors for Intelligent Quantum Sensing"

Abstract

Diamond nitrogen-vacancy (NV) centres have been developed into a leading quantum sensing platform operated in room temperature, providing the quantitative and spatially resolved detection of magnetic field, electric field and temperature through optically detected magnetic resonance (ODMR). However, the conventional wide-field sensing architecture remains a passive imager: fluorescence light is recorded frame by frame, which will be digitalised and transmitted to an external processor for fitting and further analysis. This record–transmit–post-process pipeline is slow, data-heavy and “naive”, in that it captures signals faithfully but extracts no information and especially reaches no decision by itself. These limitations, including the complicated and bulky system, the redundant transmission of raw data, and the naive and fixed sensing capability, constrain diamond quantum sensing to the laboratory and leave the fast, dynamic and edge applications unreachable. This thesis addresses these limitations by coupling diamond NV sensing within sensor and near-sensor computing, step by step moving computation/processing towards, and finally into, the sensing front-end, inspired by the biological vision system, in which the retina pre-processes information before it reaches the brain. Three contributions are presented in the thesis. This dataset supports the PhD thesis entitled "Coupling In-Sensor and In-memory Computing with Diamond Nitrogen-Vacancy Sensors for Intelligent Quantum Sensing". It contains research data generated during the design, characterization, and experimental measurement of the project. The deposited files are organized according to the relevant thesis chapters: (Dataset (Chapter 3), Dataset (Chapter 4), Dataset (Chapter 5), etc. It includes result files captured during the experiment measurement, and specifically covers the frames from EMCCD and raw data from event cameras during ODMR measurement, current readout from photosensors, characteristic data for RRAM devices, weight matrix for trained neural network and other intermediated data after processing.