Adaptive Quantization for Key Generation: Balancing Rate and Consistency

Wireless key generation systems often face challenges in achieving low Key Disagreement Rates (KDR) and Key Generation Rate (KGR) across diverse scenarios, limiting their reliability and efficiency. This paper presents a wireless key generation system based on the ESP32 platform, combining quantization methods and error correction codes to address these issues through comparative experiments. To obtain the ideal quantization order in different scenarios, we propose an adaptive quantization order selection model based on the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The results indicate that the performance of quantization methods is influenced by both the specific scenarios and the quantization order employed. Moreover, the combination of Joint Amplitude-Phase Equiprobable Quantization (JAPEQ) and Low-Density Parity-Check Codes (LDPC) outperforms other combinations in terms of KDR. Finally, our proposed adaptive quantization order selection model demonstrates ideal performance in terms of both key consistency and generation rate.