Department of Computer Science, University of York, York YO10 5GH, United Kingdom.
Phys Rev Lett. 2018 Jun 1;120(22):220505. doi: 10.1103/PhysRevLett.120.220505.
One crucial step in any quantum key distribution (QKD) scheme is parameter estimation. In a typical QKD protocol the users have to sacrifice part of their raw data to estimate the parameters of the communication channel as, for example, the error rate. This introduces a trade-off between the secret key rate and the accuracy of parameter estimation in the finite-size regime. Here we show that continuous-variable QKD is not subject to this constraint as the whole raw keys can be used for both parameter estimation and secret key generation, without compromising the security. First, we show that this property holds for measurement-device-independent (MDI) protocols, as a consequence of the fact that in a MDI protocol the correlations between Alice and Bob are postselected by the measurement performed by an untrusted relay. This result is then extended beyond the MDI framework by exploiting the fact that MDI protocols can simulate device-dependent one-way QKD with arbitrarily high precision.
在任何量子密钥分发（QKD）方案中，参数估计是至关重要的一步。在典型的 QKD 协议中，用户必须牺牲部分原始数据来估计通信信道的参数，例如误码率。这在有限尺寸的情况下在密钥率和参数估计精度之间引入了一个权衡。在这里，我们表明连续变量 QKD 不受此约束，因为整个原始密钥可同时用于参数估计和密钥生成，而不会危及安全性。首先，我们表明，由于在 MDI 协议中，Alice 和 Bob 之间的相关性是由不可信中继执行的测量进行后选择的，因此这一特性适用于测量设备无关（MDI）协议。然后，通过利用 MDI 协议可以以任意高精度模拟设备相关的单向 QKD 的事实，将该结果扩展到 MDI 框架之外。
