Evaluation of Multi-Level Amplitude and Phase Quantization for Coded Secret Key Generation

For efficient physical layer secret key generation (SKG), a key challenge is to increase the secret key rate while ensuring low inconsistency, low complexity, and low latency. This paper addresses this issue by proposing efficient amplitude and phase quantization (APQ) methods, integrated with a computationally efficient information reconciliation process. Specifically, systematic polar codes are utilized alongside a compact, universally applicable soft-bit extraction method. Furthermore, performance is optimized through the adaptive selection of quantization and information reconciliation parameters, fully leveraging the universal applicability of the proposed soft-bit estimation technique. Simulation results confirm that the proposed methods significantly reduce the initial key disagreement probability (KDP) while achieving substantial complexity reduction. In addition, the proposed methods achieve higher secret key rate with fewer parity bits.