Physical Layer Security in the Age of Artificial Intelligence and Edge Computing

Lindong Zhao,Xuguang Zhang,Jianxin Chen,Liang Zhou
DOI: https://doi.org/10.1109/mwc.001.2000044
IF: 12.777
2020-10-01
IEEE Wireless Communications
Abstract:Physical layer security (PLS) is emerging as an attractive security paradigm to complement or even replace complex cryptography. Although information-theoretical transmission optimization and physical-layer key generation have been thoroughly researched, there still exist many critical issues to be tackled before PLS is extensively applied. In this article, we investigate the prospect for exploiting artificial intelligent (AI) and edge computing (EC) to facilitate the practical application of PLS. First, two outstanding challenges facing PLS designers are identified by analyzing the fundamental assumptions regarding eavesdroppers and wireless channels. Accordingly, two enhancement schemes are designed by reaping the benefits offered by AI and EC. Specifically, a novel secure resource management framework is developed to enhance the adaptability of an optimization-based PLS paradigm, and a robust physical-layer key generation method is designed to cope with reciprocity failure. Finally, we discuss a coordinated defense architecture with multi-layer, multi-domain, and multi-dimension, which is expected to exploit the compatibility and complementarity of the existing PLS methods.
computer science, information systems,telecommunications,engineering, electrical & electronic, hardware & architecture
What problem does this paper attempt to address?
The paper aims to address the key challenges faced by Physical Layer Security (PLS) in practical applications. Specifically, the paper first reveals two main challenges in PLS design by analyzing the basic assumptions about eavesdroppers and wireless channels: 1. **Adaptability Issue**: For methods based on Signal-to-Interference-plus-Noise Ratio (SINR), these methods assume all prior information about eavesdroppers (such as location, channel state information) is known, leading to insufficient adaptability. 2. **Robustness Issue**: For methods based on reciprocity, the failure of reciprocity in wireless channel estimation results in insufficient robustness in key generation. To address these issues, the paper proposes two enhancement schemes: - For the adaptability issue of SINR methods, a security resource management framework based on active cognition is developed to enhance the adaptability of the optimization-based PLS paradigm. - For the robustness issue of reciprocity methods, a new physical layer key generation method is designed to cope with reciprocity failures. Additionally, the paper proposes a multi-layer, multi-domain, and multi-dimensional coordinated defense architecture, aiming to leverage the compatibility and complementary advantages among existing PLS methods, thereby pointing out new directions for future PLS research.