† Corresponding author. E-mail:
Project supported by the National Key Research and Development Program of China (Grant Nos. 2018YFA0306400 and 2017YFA0304100), the National Natural Science Foundation of China (Grant Nos. 61590932, 11774180, and 61705110), and the Natural Science Foundation of Jiangsu Province for Leadingedge Technology Program, China (Grant No. BK20192001).
Reference-frame-independent quantum key distribution (RFI-QKD) has been proven to be very useful and practical under realistic environment. Here, we present a scheme for one-decoy state RFI-QKD based on the work of Rusca et al. [Appl. Phys. Lett. 112, 171104 (2018)], and carry out investigation on its performance under realistic experimental conditions. Numerical simulation results show that the one-decoy state RFI-QKD can achieve comparable performance in terms of secret key rate and transmission distance as the two-decoy state correspondence under practical experimental conditions. On contrast, it does not need to prepare the vacuum state in the former case, substantially reducing the experimental complexity and random number consumptions. Therefore, our present proposal seems very promising in practical implementations of RFI-QKD.
Based on the principles of quantum mechanics, quantum key distribution (QKD) provides a way of generating information-theoretic secure keys between two distant parties, Alice and Bob. Since the first BB84 protocol[1] was proposed, much effort has been made to improve practical performance of QKD systems.[2–4] Many QKD experiments have been successfully demonstrated either in laboratories or field environments,[5–13] paving the way towards large-scale commercialization of QKD systems.
In most QKD systems, a shared reference frame is needed for Alice and Bob, which inevitably increases the cost and makes the QKD system more complicated. Fortunately, the reference-frame-independent QKD (RFI-QKD) protocol[14] was proposed by Laing et al., which can generate secret keys even with slowly drifted reference frames. Up to date, there are plenty of theoretical and experimental works were carried out on RFI-QKD.[15–21] Moreover, due to the security property of MDI protocol, MDI-QKD has attracted more and more attention recently. Combined with RFI-QKD, there are some experiments on RFI-MDI.[22,23]
On the other hand, deterministic single-photon sources are still not available, and weak coherent laser pulses are usually used in most QKD systems. It brings a chance for eavesdropper (usually called Eve) to exploit the photon-number-splitting (PNS) attack.[24–26] To combat the PNS attack, the decoy-state method[2,27–29] has been adopted in practical QKD systems, where different number of decoy states can be employed. Recently, Rusca et al.[30] did investigation on one-decoy state BB84 QKD. In their work, it is demonstrated that the one-decoy state method can show advantages for most experimental settings.
In this paper, different from Rusca et al.[30] work which focused on BB84 protocol, we extend one-decoy state method to RFI-QKD and propose a scheme of one-decoy state RFI-QKD, and further compare its performance with two-decoy state correspondence by taking statistical fluctuation analysis into account. Simulation results demonstrate that our one-decoy state RFI scheme can achieve comparable performance as two-decoy state correspondence in terms of secret key rates and transmission distances.
In the one-decoy state RFI-QKD, Alice adopts two intensities (λ ∈ {μ,v}) and three orthogonal basis (ZA, XA, and YA) to prepare quantum states, then Bob measures those states with basis ZB, XB, and YB. Generally speaking, the Z basis can be well aligned for common encoding schemes, and the X and Y basis are allowed to drift slowly with an unknown angle β. That is, the Z, X, and Y basis satisfy ZB = ZA, XB = cosβ XA + sinβYA, and YB = cosβ YA – sinβ XA. Usually, the Z basis is used to generate the final secret keys, the X and Y bases are used to estimate Eve’s information.
After the quantum communication phase, the practical detection probability and quantum-bit error-rate (QBER) with intensity λ of quantum states that Alice sends in ξA basis and Bob measures in ξB basis can be given as
Considering the post-processing speed of a QKD system mainly depends on the block size being processed, i.e., the number of pulses detected by Bob in Z basis within certain time window, here in our analysis, we assume the block size with a fixed value (nZZ). We can get the corresponding number of pulses that need to be sent (Ntot) by Alice with
Then we can get the number of detection nξAξB and the number of errors mξAξB corresponding to basis ξAξB
Through the following equation, we can calculate the fraction of nξAξB corresponding to a certain intensity λ
In the following, let us take the finite-data size effect[30,32] into account. We can obtain the relation between the observed variables
According to Ref. [30], the upper bound and the lower bound of
The lower-bound of single-photon events, the lower-bound and the upper-bound of vacuum events in basis ξAξB can be estimated with the following formulae
The upper bound of single-photon error rate
In an RFI-QKD, an intermediate quantity C can be used to monitor Eve’s information, given by
As we know, the formula on calculating the secret keys of BB84 protocol has been addressed in Refs. [30,32]
Through similar method, we can deduce out the formula on calculating the secret key rate for RFI-QKD as
In this section, we first simulate the value of C for both one-decoy and two-decoy state RFI-QKDs under different rotation angles (β), and further do comparison between their key generation rates under the same experimental conditions. Finally, we investigate the effect of the finite-data size on one-decoy RFI-QKD, we study the performance of the one-decoy state RFI-QKD with different block sizes (nZZ). It should be noted that, the efficiency of the detector and the internal losses of Bob’s apparatus are included in the value of global attenuation η in our simulations. The experimental parameters used for simulations are all listed in Table
In Fig.
Figure
In order to investigate the effect of the finite-data size on the one-decoy RFI-QKD, we compare the performance of the one-decoy state RFI-QKD with different sizes of block while setting the relative rotation angle with 20° (nZZ ∈ {105, 106, 107, 108}), as shown in Fig.
In summary, we have proposed a scheme on one-decoy state reference-frame-independent quantum key distribution, and investigate its performance under realistic environments. Numerical simulation results show that, the performance of our present scheme can exceed traditional two-decoy state RFI-QKD when the rotation angle of reference frame is not larger, e.g., < 20°. Moreover, it does not need to prepare the vacuum state in the former, and thus reduce the experimental complexity of QKD systems. Therefore, our present work could provide valuable references for practical implementation of the RFI-QKD.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] | |
[24] | |
[25] | |
[26] | |
[27] | |
[28] | |
[29] | |
[30] | |
[31] | |
[32] |