Designing Physical Layer Security Solutions For Emerging Communication Systems in 5G Networks

Abstract

Over the past few years, wireless networks have witnessed major advancements in wireless communication technologies, due to 1) the large population growth, 2) the rapid urbanization of different cities around the globe, and 3) the wide deployment of the Internet in people's daily lives. This has triggered a vigorous increase in the amount of traffic in both enterprise and residential networks, and it has motivated researchers and network operators to reconsider current network designs and mobile platforms. In order to cater for the huge expansion in the wireless industry, the 5G technology has been introduced as the next-generation standard for digital cellular networks. This technology promises vastly increased capacity, reduced latency, better utilization of resources, and faster speeds (data rates). Additionally, the 5G network architecture includes a large heterogeneous panel of interconnected networks and devices, such as Device-to-Device (D2D) and Machine-to-Machine (M2M) networks, small cell access points, network cloud, Internet of Things (IoT), and many more. One important requirement that needs to be addressed in 5G networks is its security. This is, mainly, attributed to the fact that existing security solutions and cryptographic algorithms can not support the stringent requirements of 5G networks. More specifically, conventional security solutions introduce a considerable overhead in terms of resources and delay (multi-round operations), which is not feasible for constrained devices. On the other hand, Physical Layer Security (PLS) has, recently, emerged as a promising methodology for enhancing the security of wireless networks, without relying on upper-layer cryptographic techniques. It allows legitimate users to exchange confidential messages in the presence of adversaries, by simply utilizing the dynamic properties and characteristics of wireless channels. Security in wireless networks has always been addressed separately from the physical layer, due to its uncontrollable random nature. However, with the tremendous advancement in computational capabilities, classical security techniques (static structure) are becoming less secure and the need for new adjustments is becoming more crucial. More and more research has been directed towards studying, understanding and exploiting the highly random nature of wireless networks. As a result, this has paved the way for new security solutions, that are more robust and less complex than current schemes. Moreover, the physical layer is common to all kinds of devices, hence, any security solution at this layer is useful for all heterogeneous devices. The goal of this PhD dissertation is to design and evaluate novel PLS solutions that guarantee multiple security services with minimum overhead. One important aspect is achieving a good balance between security and performance, in order to ensure the efficient and proper deployment of different state-of-the-art communication systems in 5G networks. Another aspect is providing a complete ``security framework" for emerging communication systems and resource-limited devices. This framework consists of several protocols and algorithms, that manage the transferal of information in public wireless networks. Unlike traditional security solutions, which require multiple rounds of extensive operations, the proposed PLS techniques leverage the random and dynamic properties of wireless channels to achieve robust security using a single round of simple operations. These schemes are classified according to five security services, which are: device authentication, key generation and distribution, data confidentiality, data integrity and source authentication, and data availability. For each security service, several variant schemes are presented and evaluated. The proposed security solutions target different technologies such as NOMA and MIMO. Since OFDM is expected to remain a key enabling technology in emerging and future systems such as 5G networks, PLS schemes (data confidentiality and message authentication) for OFDM systems are also designed and evaluated. Moreover, these methods are compared when applied at two instances; before the Inverse Fast Fourier Transform (IFFT) and after, since it is important to quantify the effect of each case on the security level. The security level and performance gains of the proposed schemes are analyzed using simulations and numerical results. The various obtained results prove the superiority of the proposed solutions over similar existing approaches in the literature.