Research On Performance Optimization And Realization Of Practical Measurement-device-independent Quantum Key Distribution

In recent decades,based on the quantum mechanics,some advanced technologies and industries flourished.Especially,the rapid development of quantum computer poses a great threat to conventional cryptography which is based on computational complexity.Luckily,combined with onetime-pad method,quantum key distribution(QKD)can provide an unconditional secure communication.Since the proposal of BB84 protocol in 1984,QKD has become an interdisciplinary field of study after more than 30 years' development.In future,it may shield the information security for government,finance,military and et.al.In practical implementation,however,imperfection of devices in QKD system is one of the obstructions.In 2012,researchers in Toronto University proposed a measurement-deviceindependent quantum key distribution(MDI-QKD)protocol which can be immune to all side-channel attacks directed at the detecting devices,and thus improved the security and practicality in QKD system.Nonetheless,there still exists a vast promotion roomage in performance,and the security of sources also needs careful investigation.This dissertation focuses mainly on the practical implementation of different sources in MDIQKD,where intensity-modulation error and state-preparation error are considered.Additionally,a general model of sending-or-not-sending(SNS)QKD,which contains the asymmetric case,is proposed.The research contents are as follows:1.Heralded single-photon source(HSPS)and weak coherent source(WCS)are studied in MDIQKD system accompanied with finite-size effects.Due to the extremely low vacuum state component in HSPS,the quantum bit-error rate in X base can be reduced,thus resulting a better performance in transmission distance.At the same time,modulation error is considered in three-intensity scheme,and global parameter optimization is applied.Additionally,a fourintensity scheme proposed by Zhou et.al [Phys.Rev.A 93,042324(2016)] is employed,which is combined with joint study and collective constraint method.Compared with conventional three-intensity method,this new scheme improves the final key rate and transmission distance a lot when the total number of pulses is not large.2.Perfect state preparation is required in original MDI-QKD protocol which seems to be a tough work in real-life implementation.To address this problem,researchers proposed a new method by reusing the data in mismatched basis where the states are only required to be a bidimensional state.Based on their study,a decoy-state MDI-QKD with three uncharacterized states is proposed.Thanks to the advantages in statistical fluctuation and phase error-rate estimation,our new scheme achieves better performance.In state preparation process,only one state is required in X base,and the complexity of the system is further reduced.In addition,a proof-of-principle experiment is implemented.In this experiment,polarization and bias current can be automatically adjusted for a free-running system.Based on this setup,a 170 km transmission distance is achieved,representing the longest transmission distance under the same security level on record.3.Reference-frame-independent measurement-device-independent quantum key distribution(RFI-MDI-QKD)is immune to all the detector side-channel attacks and robust against unstable reference frames at the same time.However,its practical performance suffers a lot from statistical fluctuation where only negative parameter estimations with a decoy-state method can be given.In this work,we take the collective constraint and the relationship between different bases to reduce the impact of statistical fluctuation.Furthermore,the joint study method is considered for better parameter estimation.Finally,we conduct a proof-ofprinciple demonstration of RFI-MDI-QKD and achieve a 200 km transmission distance representing the longest transmission distance with this protocol.4.Twin-field QKD garners widespread attention owing to its breakthrough of the repeaterless secure key rate bound.To fullfill the loopholes of the original protocol,SNS QKD is proposed which can tolerate large misalignment error.On this basis,a general model of SNS QKD is proposed in this work.Asymmetric case where the two users own different channels in length is included.A new constraint is adopted to fit the decoy-state method by numerical analysis.Without extra operations,the users only need to adjust the intensities separately.The simulation results indicate that this method exhibits better performance than directly applying original method after compensating the asymmetric channels.