Design of New-generation Quantum Key Distribution Systems for Secure Communication

Abstract

Quantum information processing relies on the principles of quantum physics to gain unique advantages in various information processing tasks. Among them, the quantum key distribu- tion (QKD) technique guarantees the information-theoretic security of the task of remote secure key distribution. Unlike its classical counterpart, the security of QKD does not rely on the lack of computational power of the current classical computers on one-way functions such as prime factorisation. QKD thus ensures the long-term security even at the presence of quantum computers. Moreover, QKD already enjoys mature deployments with national-wide QKD networks and satellite-assisted free-space QKD links. The performances of QKD protocols are usually judged from both the theoretical and practical sides. On the theoretical side, a QKD protocol should be secure against any possible attacks from the adversaries, with reliable key rate lower bounds giving high distillable key rate and long allowable distance. On the practical side, a QKD protocol should preferably be deployable under the standard techniques from classical communication. It should also give good performances under non-ideal settings such as device imperfection and finite data size. The current QKD protocols generally cannot triumph in both theories and practices. For instance, the Bennett-Brassard-1984 (BB84) protocol enjoys comprehensive security analysis valid under the most general attacks and finite data size, yet its single-photon detector is usually costly and requires low-temperature operation. On the other hand, the Grosshans-Grangier-2002 (GG02) protocol uses the standard coherent detection technique, yet its security is still to be completed under practical implementation. The situation is reasonable since realistic systems usually possess infinite uncharacterised dimensions, but theories are mainly effective in low dimensions. This thesis focuses on the designs of new-generation QKD protocols that are both theoretically sound and practically favourable. Specifically, three potential candidates will be discussed: • Twin-field QKD covering the longest distances but experimentally demanding • Discrete-modulated continuous-variable QKD with high practicality but incomplete in security • Time-bin-encoding continuous-variable QKD with complete security and high practi- cality Their advantages and drawbacks will be contrasted from both the theoretical and practical perspectives. Adjustments to each protocol will be proposed to gain new features yielding better practical performances. New theoretical methods will be introduced to account for the complete security of practical QKD protocols. Simulation will be performed under realistic settings illustrating the performances of these new-generation protocols and their distinct enhancements to the performances of standard QKD. The related works in this thesis are believed to point out the directions in search of the next-generation QKD protocols with robust theories and practical high performances.