Improved secure key rate for the decoy state protocol in the finite key regime

Abstract

The quantum key distribution is a promising candidate for an unconditionally secure secret communication between two parties. However, due to imperfections in the reality---more specifically, a lossy channel and the near impossibility of a single-photon generating device, its security in practical application could be undermined by various attacks, such as the photon-splitting attack. The decoy state method offers an efficient way to combat this attack, thus significantly increasing the security of quantum key distribution. Previous analyses mostly focused on three intensities, and the key rates were calculated in the asymptotic key length regime. Recently, decoy state method calculations are generalized to arbitrary number of intensities, and an efficient method of evaluation of the key rate has been devised with the use of McDiarmid inequality. In this thesis, three new methods are presented based on the further use of McDiarmid inequality. These are based on extensions of the inequality and different ways of its implementation. The Monte Carlo method is applied to numerically calculate the improved key rates. The results show a few percents of improvement compared to the old key rates. These improved key rates allow a more robust secret key communication and make the decoy-state protocol more practical.