Source Intensity Errors and Data Correction Schemes in Downstream Continuous-Variable Quantum Key Distribution Access Networks

Downstream multi-user continuous variable quantum key distribution (CV–QKD) based on passive optical networks represents a promising solution for large-scale quantum-secure access. However, global source errors can introduce biased parameter estimation and potential security vulnerabilities. To address this issue without additional hardware overhead, two data postprocessing correction schemes adapted to different deployment conditions are proposed. For scenarios with known intensity fluctuation statistics, a low-intensity pulse data tagging strategy is employed to construct conservative security bounds. For the more general scenario where only measurable intensity bounds are available, a worst-case bound-based parameter estimation framework is developed. The finite-size performance is further evaluated. Numerical results demonstrate that both schemes effectively eliminate security risks induced by intensity errors, maintain favorable multi-user scalability and acceptable finite-size performance, and provide a low-cost technical approach for the practical deployment of multi-user CV–QKD access networks.