Quantum photonic sensors array with machine learning for imaging applications

Abstract

Quantum sensors have the characteristics of high sensitivity, high accuracy and strong stability and have been used in many fields. However, though many applications require arrays of multiple quantum sensors, most of the current applications still focus on single quantum sensors. Entangled multi-quantum sensor arrays has the advantage of high precision, high accuracy, and better utilization of entanglement. In this paper, we propose an entangled photonic quantum sensor network with a radio frequency spectral range with the sensitivity and accuracy of sensors even beyond the standard quantum limit (SQL). Based on this model, we derived the variance model of the sensor array and analyzed the optimize parameter settings the sensor array and verified it by simulation. Entangled sensor network performs much better than classical sensor structure, but entanglement also increase the complexity for probe state setting. Then, we utilize sensor arrays to accomplish supervised learning (SL) tasks. We introduce supervised learning assisted by entangled sensor networks (SLAEN) to perform SL tasks at the physical layer. The entanglement shared by sensors in SLAEN improves the performance of extracting global features of the object under investigation. We leverage SLAEN to build an entanglement-assisted support vector machine for data classification, conclude that SLAEN has a considerable improvement in entanglement performance compared to traditional strategies, which is around 10 times more accuracy than traditional model when no loss applied and still maintaining 2 times lower error probability when suffered from experimental imperfections. With only one photon and 10% loss applied, SLAEN model can reach 0.05 error probability. Thus, we reach a conclusion that SLANE is a good way to to find optimal entanglement probe states and measurement settings to maximize entanglement functionality.