Physical layer security schemes for wireless communications are currently crossing the chasm from theory to practice. They promise information-theoretical security, for instance by guaranteeing the confidentiality of wireless transmissions. Examples include schemes utilizing artificial interference---that is 'jamming for good'---to enable secure physical layer key exchange or other security mechanisms. However, only little attention has been payed to adjusting the employed adversary models during this transition from theory to practice. Typical assumptions give the adversary antenna configurations and transceiver capabilities similar to all other nodes: single antenna eavesdroppers are the norm. We argue that these assumptions are perilous and 'invite the thief'. In this work, we evaluate the security of a representative practical physical layer security scheme, which employs artificial interference to secure physical layer key exchange. Departing from the standard single-antenna eavesdropper, we utilize a more realistic multi-antenna eavesdropper and propose a novel approach that detects artificial interferences. This facilitates a practical attack, effectively 'lockpicking' the key exchange by exploiting the diversity of the jammed signals. Using simulation and real-world software-defined radio (SDR) experimentation, we quantify the impact of increasingly strong adversaries. We show that our approach reduces the secrecy capacity of the scheme by up to 97% compared to single-antenna eavesdroppers. Our results demonstrate the risk unrealistic adversary models pose in current practical physical layer security schemes.
Physical layer security schemes for wireless communications are currently crossing the chasm from theory to practice. They promise information-theoretical security, for instance by guaranteeing the confidentiality of wireless transmissions. Examples include schemes utilizing artificial interference---that is 'jamming for good'---to enable secure physical layer key exchange or other security mechanisms. However, only little attention has been payed to adjusting the employed adversary models during this transition from theory to practice. Typical assumptions give the adversary antenna configurations and transceiver capabilities similar to all other nodes: single antenna eavesdroppers are the norm. We argue that these assumptions are perilous and 'invite the thief'. In this work, we evaluate the security of a representative practical physical layer security scheme, which employs artificial interference to secure physical layer key exchange. Departing from the standard single-antenna eavesdropper, we utilize a more realistic multi-antenna eavesdropper and propose a novel approach that detects artificial interferences. This facilitates a practical attack, effectively 'lockpicking' the key exchange by exploiting the diversity of the jammed signals. Using simulation and real-world software-defined radio (SDR) experimentation, we quantify the impact of increasingly strong adversaries. We show that our approach reduces the secrecy capacity of the scheme by up to 97% compared to single-antenna eavesdroppers. Our results demonstrate the risk unrealistic adversary models pose in current practical physical layer security schemes.
Physical layer security for wireless communication is broadly considered as a promising approach to protect data confidentiality against eavesdroppers. However, despite its ample theoretical foundation, the transition to practical implementations of physical-layer security still lacks success. A close inspection of proven vulnerable physical-layer security designs reveals that the flaws are usually overlooked when the scheme is only evaluated against an inferior, single-antenna eavesdropper. Meanwhile, the attacks exposing vulnerabilities often lack theoretical justification. To reduce the gap between theory and practice, we posit that a physical-layer security scheme must be studied under multiple adversarial models to fully grasp its security strength. In this regard, we evaluate a specific physical-layer security scheme, i.e. orthogonal blinding, under multiple eavesdropper settings. We further propose a practical "ciphertext-only attack" that allows eavesdroppers to recover the original message by exploiting the low entropy fields in wireless packets. By means of simulation, we are able to reduce the symbol error rate at an eavesdropper below 1% using only the eavesdropper's receiving data and a general knowledge about the format of the wireless packets.
The wireless channel is inherently an open medium which exposes wireless communication users to the risk of interception. The vulnerabilities of the wireless channel have led to the search for innovative and efficient security solutions at different levels of the network protocol. In particular, the physical layer level allows for implementing some security functions, such as key generation and exchange, by exploiting the intrinsic randomness of the wireless communication channel. In this paper, we analyze the effect of wireless channel and of antenna characteristics on physical layer security (PLS) key generation. The aim is to put in evidence the relation between physical quantities, such as radio channel characteristics or antenna directivity, with higher level secrecy performance indicators, such as the key mismatch between the different communication nodes and the key randomness. For this purpose, we developed a comprehensive numerical model, which implements all the different steps, from signal generation and wireless propagation modeling to key generation through spectral quantization. The reported results show that the multipath richness of the radio channel, as in outdoor urban scenarios, improves the PLS secrecy performances with respect to simpler propagation scenarios. Moreover, we show that the use of more directive antennas in the legitimate communication link improves secrecy. An opposite effect is achieved when the directivity of the eavesdropper antenna is increased, thus improving its ability to perform passive attacks.
Achieving secure transmission in wireless network is a significantly important research topic. Classic cryptography encrypts a confidential message to an unreadable cipher message. Thus only the authentic receiver with correct secrecy key is able to decrypt the message. However, with the growing computational capability of adversaries, the requirement of secrecy key generation and distribution is getting stringent. Another method to achieve wireless security, which has attracted much attention in recent years, is through physical layer processing, named physical layer security. In a typical wiretap channel model consisting of a transmitter, a legal receiver and an adversary (illegal receiver), early works have proven that it is possible to achieve positive secrecy rate using multiple antenna techniques, such as beamforming, at the transmitter. In this thesis, we focus on two types of malicious behaviors of the adversary: passive eavesdropping and active attack. For the passive eavesdropping, we first consider the multi-input-single-output (MISO) wiretap channel, in which the legitimate transmitter utilizes the artificial noise aided precoding strategy to maximize the achievable secrecy rate of the channel. Both scenarios of known and unknown channel state information (CSI) for the eavesdropper channels are considered. By deriving the closed-form expression of ergodic secrecy rate, we prove that there exists an optimal power ratio between the information signal and the artificial noise. We then extend our study to a four-node fading channel model in a wireless relay network, which includes two single-antenna users, one single-antenna eavesdropper and one relay node equipped with multiple antennas. By studying the decode-and-forward (DF) scheme, cooperative jamming (CJ) scheme and one proposed hybrid relay scheme, relay-and-jamming (RJ), we aims to find the optimal relay scheme with the condition that the transmitter and the relay node are under individual power constraints and the CSI of the illegitimate channel is only statistically known. Our analysis shows that the hybrid RJ scheme is optimal when the transmitter has a relatively small power budget and/or the transmitter is far away from the intended receiver. Besides the passive eavesdropping, the adversary could also choose the active attack. The pilot spoofing attack is one active attack conducted by the adversary during the channel estimation phase of the legitimate transmission. In practical systems, the pilot signals are usually known \textit{a priori} and repeatedly used information. In this attack, an intelligent adversary spoofs the transmitter on the estimation of CSI by sending the identical pilot signal as the legitimate receiver. By doing so, the adversary could obtain a larger channel rate in the data transmission phase, and also drastically weaken the strength of the received signal at the legitimate receiver if the adversary utilizes large enough power. Due to the serious consequences caused by such an attack, we first propose an energy ratio detector (ERD) that explores the asymmetry of received signal power levels at the transmitter and the legitimate receiver when the attack occurs. Our analysis shows that the ERD could provide very high probability of detection with certain requirement on the false alarming probability. Furthermore, we design a two-way training based scheme to provide not only attack detection but also secure transmission. The two-way training based detector (TWTD) exploits the intrusive component created by the adversary, followed by a secure beamforming-assisted data transmission. In addition to the good detection performance, our scheme could estimate both legitimate and illegitimate channels, with which the beamforming could be designed for data transmission.
Wireless communications offer data transmission services anywhere and anytime, but with the inevitable cost of introducing major security vulnerabilities. Indeed, an eavesdropper can overhear a message conveyed over the open insecure wireless media putting at risk the confidentiality of the wireless users. Currently, the way to partially prevent eavesdropping attacks is by ciphering the information between the authorised parties through complex cryptographic algorithms. Cryptography operates in the upper layers of the communication model, bit it does not address the security problem where the attack is suffered: at the transmission level. In this context, physical layer security has emerged as a promising framework to prevent eavesdropping attacks at the transmission level. Physical layer security is based on information-theoretic concepts and exploits the randomness and the uniqueness of the wireless channel. In this context, this thesis presents signal processing techniques to secure wireless networks at the physical layer by optimising the use of multiple-antennas. A masked transmission strategy is used to steer the confidential information towards the intended receiver, and, at the same time, broadcast an interfering signal to confuse unknown eavesdroppers. This thesis considers practical issues in multiple-antenna networks such as limited transmission resources and the lack of accurate information between the authorised transmission parties. The worst-case for the security, that occurs when a powerful eavesdropper takes advantage of any opportunity to put at risk the transmission confidentiality, is addressed. The techniques introduced improve the security by offering efficient and innovative transmission solutions to lock the communication at the physical layer. Notably, these transmission mechanisms strike a balance between confidentiality and quality to satisfy the practical requirements of modern wireless networks.
The current research on physical layer security is far from implementations in practical networks, arguably due to impractical assumptions in the literature and the limited applicability of physical layer security. Aiming to reduce the gap between theory and practice, this thesis focuses on wireless physical layer security towards practical assumptions and requirements. In the first half of the thesis, we reduce the dependence of physical layer security on impractical assumptions. The secrecy enhancements and analysis based on impractical assumptions cannot lead to any true guarantee of secrecy in practical networks. The current study of physical layer security was often based on the idealized assumption of perfect channel knowledge on both legitimate users and eavesdroppers. We study the impact of channel estimation errors on secure transmission designs. We investigate the practical scenarios where both the transmitter and the receiver have imperfect channel state information (CSI). Our results show how the optimal transmission design and the achievable throughput vary with the amount of knowledge on the eavesdropper's channel. Apart from the assumption of perfect CSI, the analysis of physical layer security often ideally assumed the number of eavesdropper antennas to be known. We develop an innovative approach to study secure communication systems without knowing the number of eavesdropper antennas by introducing the concept of spatial constraint into physical layer security. That is, the eavesdropper is assumed to have a limited spatial region to place (possibly an infinite number of) antennas. We show that a non-zero secrecy rate is achievable with the help of a friendly jammer, even if the eavesdropper places an infinite number of antennas in its spatial region. In the second half of the thesis, we improve the applicability of physical layer security. The current physical layer security techniques to achieve confidential broadcasting were limited to application in single-cell systems. The primary challenge to achieve confidential broadcasting in the multi-cell network is to deal with not only the inter-cell but also the intra-cell information leakage and interference. To tackle this challenge, we design linear precoders performing confidential broadcasting in multi-cell networks. We optimize the precoder designs to maximize the secrecy sum rate with based on the large-system analysis. Finally, we improve the applicability of physical layer security from a fundamental aspect. The analysis of physical layer security based on the existing secrecy metric was often not applicable in practical networks. We propose new metrics for evaluating the secrecy of transmissions over fading channels to address the practical limitations of using existing secrecy metrics for such evaluations. The first metric establishes a link between the concept of secrecy outage and the eavesdropper's ability to decode confidential messages. The second metric provides an error-probability-based secrecy metric which is often used for the practical implementation of secure wireless systems. The third metric characterizes how much or how fast the confidential information is leaked to the eavesdropper. We show that the proposed secrecy metrics enable one to appropriately design secure communication systems with different views on how secrecy is measured.
Physical layer security schemes for wireless communication systems have been broadly studied from an information theory point of view. In contrast, there is a dearth of attack methodologies to analyze the achievable security on the physical layer. To address this issue, we develop a novel attack model for physical layer security schemes, which is the equivalent to known-plaintext attacks in cryptoanalysis. In particular, we concentrate on analyzing the security of orthogonal blinding schemes that disturb an eavesdropper’s signal reception using artificial noise transmission. We discuss the theory underlying our attack methodology and develop an adaptive filter trained by known-plaintext symbols to degrade the secrecy of orthogonal blinding. By means of simulation and measurements on real wireless channels using software-defined radios with OFDM transceivers, we obtain the operating area of our attack and evaluate the achievable secrecy degradation. We are able to reduce the secrecy of orthogonal blinding schemes to Symbol Error Rates (SERs) below 10% at an eavesdropper, with a knowledge of only a 3% of the symbols transmitted in typical WLAN frames.