Shihong Pan

Detector noise is a critical factor in determining the performance of a quantum key distribution (QKD) system. In continuous-variable (CV) QKD with optical coherent detection, the trusted detector noise model is widely used to enhance both the secret key rate and transmission distance. This model assumes that noise from the coherent detector is inherently random and cannot be accessed or manipulated by an adversary. Its validity rests on two key assumptions: (1) the detector can be accurately calibrated by the legitimate user and remains isolated from the adversary, and (2) the detector noise is truly random. So far, extensive research has focused on detector calibration and countermeasures against detector side-channel attacks. However, there is no strong evidence supporting assumption (2). In this paper, we analyze the electrical noise of a commercial balanced Photoreceiver, which has been applied in CV-QKD implementations, and demonstrate that assumption (2) is unjustified. To address this issue, we propose a “calibrated detector noise” model for CV-QKD, which relies solely on assumption (1). Numerical simulations comparing different noise models indicate that the new model can achieve a secret key rate comparable to the trusted-noise model, without depending on the questionable assumption of “truly random” detector noise.

The coherent one-way (COW) protocol is a promising commercial solution to practical quantum key distribution (QKD) due to its simple optical implementation. However, the non-IID structure of COW due to its inter-signal coherence makes standard security analysis inapplicable. Recently, it has been shown that a modified COW setup allows standard IID analysis, but at the cost of imposing extra limitations and increasing the number of pulses required for each bit. Here we propose a variant that possesses the IID structure and completely retains the optical setup of COW, but with a different data processing scheme that ignores inter-signal information. We obtain key rate lower bound close to analysis for the previously proposed IID variant, and achieves a higher number of key bits transmitted per second.