Recent progress in wireless system and network security exploits wireless channel characteristics, also known as link signatures, to offer enhanced and sometimes the only available security solutions for ubiquitous wireless applications; two prominent examples are link-signature based secret key generation and physical layer authentication. However, it is found that the current link-signature study makes wrong assumptions about the adversary's inference capability, which will expose the legitimate systems to severe threats with little awareness.
Objective: The objective of this project is to seek a fundamental understanding of the limits in adversarial inference of wireless channel characteristics, and develop stronger link-signature based security mechanisms for wireless communications. Various analytical channel correlation models involving important physical propagation factors in typical wireless environments will be investigated, and effective inference algorithms and practical designs will be explored from the adversary's position, to determine the vulnerability of existing wireless security mechanisms. Then, two approaches will be considered to further strengthen the current wireless security solutions: one is to exploit multiple-antenna and cooperative communication techniques to disrupt the adversarial inference process, and the other is to integrate cryptographic techniques into link-signature constructions to achieve improved and guaranteed protection.
Intellectual merit: The intellectual merit is correction of a wide-spread misconception about the link-signature security, and establishment of a proven scientific foundation facilitating the development of new wireless security mechanisms.
Broader impacts: The broader impacts are significantly improvement in the security of current and emerging wireless devices and systems, cross-disciplinary education, and well-rounded training of future IT workforce.
