The Improvement of Performance for Continuous-Variable Quantum Key Distribution with Imperfect Gaussian Modulation

Gaussian modulation is one of the key steps for the implementation of continuous-variable quantum key distribution (CVQKD) schemes. However, imperfection in the Gaussian modulation may introduce modulation noise that can deteriorate the performance of CVQKD systems. In this paper, we mainly investigate how to improve the performance of a CVQKD system from different aspects. First, we explore the several different origins, impacts and monitoring schemes for the modulation noise in detail. Then, we discuss the practical performance of a CVQKD system with an untrusted noise model and neutral party model, respectively. These analyses indicate that the neutral party model should be reasonably regarded as a general noise model, which will passively and greatly raise the performance of the system. Further, we propose a dynamic auto-bias control scheme to actively resist the modulation noise which comes from the drift of bias point of the amplitude modulator. Together these methods contribute to the improvement of the practical performance of CVQKD systems with imperfect Gaussian modulation.