Abstract
Cryptography is used to protect sensitive and classified information in technologyand science areas such as cryptocurrency Bitcoin and cryptography assurance
in Blockchain. Bitcoin is the first decentralized cryptocurrency that had attracted
a rapid growth both in value and popularity since its first release in late 2008. Its
security is based on cryptographic protocols that allow users to perform secure online
transactions. These protocols include digital signatures and encryption/decryption
algorithms based on elliptic curves. The security of Bitcoin transactions is mainly
relied on a random binary sequence generated from a random number generator and
a private key to produce digital signatures that are used to validate the authenticity
and integrity of Bitcoins. However, Bitcoin is a relatively new technology and has
security challenges related to current key generation algorithms and random number
generators. Recently, biometrics have been investigated for use in cryptographic
applications such as key generation for encryption and decryption algorithms. In
general, people can use their biometrics (including voice, face, iris, and fingerprint)
as good alternatives, or supplements, to PINs and passwords for generating keys. In
recent years, the brainwave signal has been investigated for person identification and
authentication because of its benefits over traditional means, and it has been shown
that the brainwave signal can be a good alternative for these cryptographic applications.
However, little work has been done on the combination of cryptography and
brainwave signals.
This research project focuses on the use of Electroencephalography (EEG) as
a new biometric to build EEG-based cryptosystems. These cryptosystems include
EEG-based cryptographic key generation systems using given EEG signals with an
authentication mechanism to generate keys that can be applied in the protection
of Bitcoin wallets. The key generation systems are based on the quasi-stationary
characteristics of the EEG that is used for person identification and authentication.
This research also covers EEG-based random number generators based on EEG nonlinear
and chaotic characteristics to enhance the security of digital signatures used in
Bitcoin transactions. In addition, this research investigates the impacts of epilepsy,
alcohol, and emotion characteristics on these systems, matters that have not been
investigated and evaluated in previous studies.
Evaluation experiments performed on the Australian EEG, Alcoholism, GrazIIIa,
GrazA 2008 and DEAP datasets show better results for most of the proposed
techniques than current traditional techniques to have lower values of false acceptance
rates and false rejection rates. The experimental results also achieve higher key
length generated, a factor that is crucial in practical security applications to make
revealation of the keys more difficult. The experimental results have also shown
that epilepsy, alcohol and emotion characteristics have impacts on the EEG-based
key generation systems, but they do not affect the length and the randomess of the
generated key. Moreover, these characteristics do not affect on EEG-based random
number generators.
| Date of Award | 2019 |
|---|---|
| Original language | English |
| Supervisor | Dat TRAN (Supervisor) & Dharmendra Sharma AM PhD (Supervisor) |