Candidate: Seyedehsan Bateni
Title: Wireless One-time PAD
Date: February 26, 2018
Time: 10:40am
Place: EIT 3142
Supervisor(s): Khandani, Amir K.
Abstract:
Most commonly used security methods rely on cryptographic techniques employed at the upper layers of a wireless network. These methods basically rely on the computational hardness of some mathematical problems. This computational complexity nature may not hold in near future due to rapid development of hardware technology, yet not taking into account the revolution of Quantom computing in further future. Another core problematic issue exists in symmetric security systems; there is a deadlock between securing the channel and establishing the shared key. We need the key for securing the channel, on the other hand for sharing the key we need a secure channel. It is obvious that there is a loop here.
To overcome such vulnerabilities, physical layer security (PLS) has been widely studied recently. PLS schemes build on the idea of turning unpredictable and random wireless channel characteristics into a source for information-theoretic security. Information-theoretic security itself, relies on Shannon's pioneer work . Shannon, inspired by one-time pad, also known as Vernam cipher, theoretically showed that the only unconditional perfect secrecy system is a one-time pad with a key at least as random as the plaintext, i.e , a system that uses a different random key to cipher any new plaintext. In PLS key generation methods, legitimate parties alternatively send probe signals and estimate CSI (Channel State Information) of common random channel and then convert enough amount of these estimates to bit strings as shared keys. To achieve perfect secrecy, the key generation methods must meet the key randomness and key generation rate (KGR) requirements of their specific cryptographic applications.
In this research a novel practical system for achieving unconditional perfect secrecy is presented. Our system uses channel-phase as the probing parameter to fully benefit from its uniform distribution over [0,2*pi]. Moreover, by intentionally perturbing the wireless channel in vicinity of the transceiver antenna based on RF mirrors structure, it produces different random phase values for each channel probing much faster than the inherent channel variation would do, resulting in dramatically higher KGR than any wide-band PLS scheme presented so far and realizing true perfect secrecy. Last but not least, a detailed practical algorithm for implementing the system as well as empirical results is the main focus of this research which makes our system the first channel-phase-based PLS scheme implementation, reported so far.