The objective of this project is to demonstrate how electrically reconfigurable antennas can be coupled with physical layer based security algorithms to significantly augment wireless network security. The large scale proliferation of wireless technology brings with it a more serious concern for security since wireless data transmission introduces multiple avenues for attack and penetration into a network. Reconfigurable antennas are antennas that can be electrically modified to exhibit different radiation patterns. The approach is to develop a framework for wireless network security based on the capabilities of reconfigurable antennas and techniques adapted from wireless communications, signal processing, and information theory.

The intellectual merit of this project is to address the technical challenges associated with wireless network security by using reconfigurable antennas. The ability of reconfigurable antennas to generate different channel realizations from a single physical antenna also has the potential to greatly augment physical layer based security algorithms and encryption-key generation techniques. Emphasis will be placed on the development of compact reconfigurable antenna structures, fingerprints for device identification, and encryption key generation based on reconfigurable antennas.

This project will have broader impacts beyond the stated research goals, as the research component of this project will be tightly integrated with an educational component that seeks to encourage design and innovation among undergraduate students as well as professionals seeking continuing education. Drexel University is a Center of Excellence for Information Assurance Education. This Center will be augmented by course modules in physical layer security and demonstrations with actual hardware.

Project Report

Businesses and homes are increasingly adopting wireless networks for the last-mile portion of their network infrastructure. Such a large scale proliferation of wireless technology, brings with it a more serious concern for security since wireless data transmission introduces multiple avenues for attack and penetration into a network. These risks will continue to increase in number and sophistication as wireless networks start to carry increasingly more sensitive information. While strong cryptographic techniques have been developed and employed to protect against such attacks, they have been shown to have weaknesses and be susceptible to failure in relatively quick time. In addition, concerns have been raised regarding the strength of the keys used in state-of-the-art cryptography. The objective of this proposal was to demonstrate how electrically reconfigurable antennas can be coupled with physical layer based security algorithms to significantly augment wireless network security by using the unique properties of the wireless channel between the communicating users. Reconfigurable antennas are antennas that can be electrically modified to exhibit different radiation patterns and/or polarizations. The ability to develop and use reconfigurable antennas for potentially new applications like wireless security radically changes the direction in which these antennas can be developed and deployed in large-scale networks. For this proposal, the PIs investigated the effect of augmenting the wireless network security using reconfigurable antennas by studying three security principals: user authentication, encryption, and intruder mitigation. User authentication and encryption is widely accepted and used throughout the world by implementing passwords, which are highly vulnerable to eavesdropping attacks. This proposal took a Physical (PHY) layer based approach in securing the wireless communications. Rather than utilizing secret keys and authentication algorithms run at application layer, the PIs explored the unique characteristics of the wireless PHY layer. For each wireless transmitter and receiver pair, a unique and reciprocal link exist in the wireless medium. This link yields a distinct channel fingerprint that changes based on the environment (e.g. the architecture of the room, movement of people) and location of the wireless nodes. Other studies have shown that leveraging this channel fingerprint can be used as a way to improve wireless network security. To start, the PIs have investigated PHY layer based user authentication and encryption techniques. The channel fingerprint was shown to change based on location of the nodes. Hence, after completing a training sequence, software defined radio (SDR) nodes were able to determine if the transmitting node is authentic or not within a defined confidence range. This decision was made by using Generalized Likelihood Ratio Test (GLRT) and Generalized Method of Moments (GMM) tests. In both scenarios, the additional degree of freedom added into the system by the use of reconfigurable antennas proved that a better decision (higher detection rates, and lower false alarm rates) is provided in terms of user authentication. This agreed with the hypothesis that was being tested: Reconfigurable antennas yield multiple realizations of the channel by leveraging their pattern diversity. These realizations deliver a stronger knowledge of the transmitter’s wireless link and, hence, improve the user authentication performance for wireless networks. The PIs continued with investigating how the unique characteristic of the wireless PHY layer channel fingerprint can be used in conjunction with reconfigurable antennas in order to improve the current encryption techniques. An algorithm was developed that leverages normalizing, averaging, and thresholding in order to extract secret bits from the channel fingerprint. For the purposes of this stage, it should be kept in mind that a theoretic wireless channel is reciprocal between the transmitter and the receiver. Expanding on this knowledge, both sides of the communication should be able to extract the same secret bits out of their fingerprint. This is vital because it facilitates agreeing on an encryption key without exchanging any information related to the key; therefore, protecting the communication from an eavesdropping attack. The PIs performed experiments that show an outside intruder is not able to gain enough knowledge regarding the key being used by trying to get close to the receiver. The use of reconfigurable antennas was also shown to improve the randomness and the length of the keys extracted compared to a standard omnidirectional antenna. Last, the directionality feature of the reconfigurable antennas were explored in order to implement intruder mitigation techniques. A radio interference module was built and integrated into an SDR, where an online scanning was involved to determine the direction of the intruder. The algorithm was designed to then automatically cause interference in the direction of the intruder and therefore, disabling him from hearing and information. The PIs for this proposal have successfully demonstrated that reconfigurable antennas are a strong addition to wireless PHY layer security techniques, specifically user authentication, encryption, and intruder mitigation. This opens up possibilities for these techniques to be implemented in COTS devices and also for cross-layer security implementations.

Project Start
Project End
Budget Start
2010-09-15
Budget End
2014-08-31
Support Year
Fiscal Year
2010
Total Cost
$359,506
Indirect Cost
Name
Drexel University
Department
Type
DUNS #
City
Philadelphia
State
PA
Country
United States
Zip Code
19102