Efficient physical layer key generation technique in wireless communications

Rushan Lin,Li Xu,He Fang,Chuan Huang

doi:10.1186/s13638-019-1634-7

Abstract

Wireless communications between two devices can be protected by secret keys. However, existing key generation schemes suffer from the high bit disagreement rate and low bit generation rate. In this paper, we propose an efficient physical layer key generation scheme by exploring the Received Signal Strength (RSS) of signals. In order to reduce the high mismatch rate of the measurements and to increase the key generation rate, a pair of transmitter and receiver separately apply adaptive quantization algorithm for quantifying the measurements. Then, we implement a randomness extractor to further increase key generation rate and ensure randomness of generated of keys. Several real-world experiments are implemented to verify the effectiveness of the proposed scheme. The results show that compared with the other related schemes, our scheme performs better in bit generation rate, bit disagreement rate, and randomness.

Highlights

In order to guarantee the confidentiality, integrity, and authenticity of communication, secret keys should be established to protect the wireless network
There have been many research works for key extraction based on Received Signal Strength (RSS) [10,11,12,13,14,15], existing methods may suffer from low bit generation rate (BGR) and high bit disagreement rate (BDR)
We propose a modified channel quantization with single guard-band (MCQSG) algorithm, which divides the quantization guard-band by different conditions for each interval

Summary

Introduction

In order to guarantee the confidentiality, integrity, and authenticity of communication, secret keys should be established to protect the wireless network. This scheme was able to achieve 0-bit disagreement, but the bit generation rate becomes too low They proposed an n-dimension quantization method to reuse each RSS n times. Alice and Bob continuously collect RSS measurements by repeatedly sending probe signals Small-scale fluctuations are reduced based on wavelet shrinkage, which is mainly caused by the divergences between Alice and Bob’s measurements.

Quantization and encoding—an adaptively MCQSG algorithm

Performance evaluation

Conclusion