An Electrocardiogram-Based Authentication Protocol In Wireless Body Area Network

Ramli, Sofia Najwa (2016) An Electrocardiogram-Based Authentication Protocol In Wireless Body Area Network. Doctoral thesis, Universiti Teknikal Malaysia Melaka.

[img] Text (24 Pages)
An Electrocardiogram-Based Authentication Protocol In Wireless Body Area Network 24 Pages.pdf - Submitted Version

Download (393kB)
[img] Text (Full Text)
An Electrocardiogram-Based Authentication Protocol In Wireless Body Area Network.pdf - Submitted Version
Restricted to Registered users only

Download (2MB)

Abstract

In the past few years, the applications of Wireless Body Area Network (WBAN) have improved the ability of healthcare providers to deliver appropriate treatments to the patients either in hospitals or at homes. Precisely, biomedical sensors in a WBAN collect physiological signal from human’s body to enable remote, continuous and real-time network services. As the signal contains highly sensitive medical information about the patient and communicates through an open wireless environment, securing the information from unauthorized access and tampering are critically needed. One of the most crucial components to support security architecture in WBAN is its key management as it serves as the fundamental of authentication and encryption, but the overheads are enormous in dealing with key generation, exchange, storage and replacement. In response to such issue, the most promising solution for key management is the use of biometrics so that the involved parties can agree on a key to provide the authenticity of medical data in WBAN. However, the existing models are inappropriate to achieve optimal security performance and the required lightweight manners due to the sensor’s resource constraints in terms of power consumption and memory space. Therefore, this thesis presents a new authentication protocol model that utilizes Electrocardiogram (ECG) signal as biometric as well as cryptographic key to ensure that the transmitted data are originated from the required WBAN. The proposed model is developed and simulated on Matlab based on an improved fuzzy vault scheme with a lightweight error correction algorithm to reduce the computational complexity when compared to previous work. To validate the proposed ECG-based authentication protocol model, the FAR and FRR analysis is done and then followed by the complexity analysis. The result of FAR and FRR analysis demonstrates that choosing a definite degree and tolerance level can achieve optimal security performance required in WBAN communications. In complexity analysis, based on t-test, the result shows that there is a significant difference with 5% significant level in the computational complexity between the proposed authentication model and the previous protocol called ECG-IJS scheme and the proposed model requires fewer overheads in terms storage and communication overheads. To enhance the overall performance, this thesis also evaluates the uniqueness and the stability of ECG signal using Independent Component Analysis (ICA) and fast Fourier Transform (FFT) algorithm respectively as the signal is applied as inputs of the proposed ECG-based authentication protocol model. The experimental result of ICA algorithm exhibits that each ECG signal is unique to each other as each signal is composed strongly from each different independent component and approximately zero relative to other independent components. While the result of FFT algorithm summarizes that the number of the common FFT peak location index for sensors on the same subject is significantly higher compared to the number of common feature for sensors on different subjects.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: Source separation (Signal processing), Signal processing, Digital techniques, Electrocardiography, Electrocardiogram, Wireless, Network
Subjects: T Technology > T Technology (General)
T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Library > Tesis > FTMK
Depositing User: Nor Aini Md. Jali
Date Deposited: 22 May 2017 00:37
Last Modified: 08 Oct 2021 15:10
URI: http://eprints.utem.edu.my/id/eprint/18534
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item