基于设备与信道特征的物理层安全方法

接入安全与数据保密是无线网络安全性和保密性的两个最重要的因素.然而,基于计算安全的身份认证及保密通信方法在未来信息化系统中面临巨大挑战.与此同时,基于信息论安全的物理层安全为身份认证和保密通信开辟了新的思路.本文综述了近年来基于设备与信道特征的物理层安全方法的研究进展.利用无线通信设备、信道的特性可以从物理层实现设备身份的识别与认证以及密钥的分发与更新,同时具备高度安全性与使用便捷性.其中,设备指纹方法从发射信号中提取发送设备的特征,作为设备身份的唯一标识,从而准确识别不同发射源个体.指纹的唯一性、鲁棒性、长时不变性、独立性、统一性和可移植性是设备指纹身份认证的依据.而基于信道特征的密钥生成方法则从接收射频信号中提取互易的上下行信道的参数,转化为对称密钥,实现一次一密的安全传输.同样地,密钥的一致性、随机性、防窃听性则是反映无线信道密钥生成方法性能的关键要素.本文对设备指纹与信道密钥的关键要素归纳分析,并指出目前存在的几类难点问题.最后,本文讨论了在未来移动通信中该技术新的应用场景.

Research on Physical-layer Security Based on Device and Channel Characteristics

Access security and data confidentiality are two of the most critical factors in wireless network security. However, existing authentication and secure transmission methods based on computation security face challenges in the future information systems. Meanwhile, physical-layer security based on information-theoretical security inaugurates a new idea to authentication and secure transmission. This paper summarizes the research progress of physical layer security methods based on device and channel characteristics in recent years. In the physical layer, the characteristics of wireless communication devices and channels can be used to identify/authenticate device identities, as well as distribute and update keys. They are both highly secure and very easy to deploy in practical applications. Device fingerprinting extracts features from the emitted signals as the unique identification for authenticated users. Uniqueness, robustness, long-term invariance, independence, consistency, and portability are six certification basises of the device fingerprinting. Secret key generation based on channel characteristics extracts reciprocal channel parameters from received ratio signals, and turn them into a pair of symmetric keys to realize one-time-pad secure communications. Similarly, key consistency, randomness, and anti-eavesdrop are crucial elements that reflect the performance of the wireless channel key generation method. This paper analyses and summarizes the critical issues of device fingerprinting and secret key generation. Finally, this paper discusses the research difficulties and new application scenarios in the future.