Sistem Autentikasi Biometrik berbasis Fitur Spektrum Sinyal Elektroensefalografi

Gaol, Jeff L. (2018) Sistem Autentikasi Biometrik berbasis Fitur Spektrum Sinyal Elektroensefalografi. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
07111440000080-Undergraduate_Theses.pdf - Published Version
Restricted to Repository staff only

Download (12MB) | Request a copy

Abstract

Sistem autentikasi terkini umumnya memiliki kelemahan, antara lain kehilangan kartu dan kunci, dan rentannya keamanan saat memasukkan kata sandi dan Personal Identification Number. Sistem autentikasi biometrik, seperti sidik jari dan retina, juga memiliki kelemahan, yaitu dapat ditiru oleh penipu. Pada penelitian ini, akan digunakan fitur sinyal elektroensefalografi (EEG) sebagai sebuah modalitas baru dalam autentikasi biometrik, karena sifatnya yang tidak sadar dan tidak dapat diimitasi dengan artefak yang bukan organisme hidup. Konfigurasi elektroda yang digunakan adalah dengan Fp2 untuk sinyal masukan, A1 untuk referensi, dan A2 untuk common-mode. Elektroda yang digunakan adalah elektroda Ag/AgCl sekali pakai. Perangkat keras instrumentasi sinyal EEG terdiri dari filter radio frequency interference, rangkaian proteksi, penguat instrumentasi, rangkaian common-mode rejection, rangkaian penghilang tegangan DC diferensial, rangkaian penguat tak membalik, low-pass filter dengan frekuensi cut-off 72Hz, high-pass filter dengan frekuensi cut-off 0,23Hz, notch filter dengan frekuensi resonansi 50Hz, rangkaian isolasi, adder, dan dikonversi ke digital dengan ADS1115. Total penguatan yang diberikan oleh rangkaian ini adalah 30.375,62x. Sinyal digital kemudian dikirim ke Arduino Nano dan diproses menggunakan Personal Computer untuk pengolahan sinyal EEG. Sinyal EEG kemudian difilter secara digital dengan filter Butterworth orde 3 dengan frekuensi cut-off 4-14Hz, diberi window Hamming, dianalisis spektrum frekuensinya menggunakan Fast Fourier Transform, dikelompokkan dalam 128 bin, dan dinormalisasi. Fitur disimpan menggunakan Microsoft SQL Server dan dikenali oleh Jaringan Syaraf Tiruan. Persentase keberhasilan verifikasi sistem ini mencapai 96% untuk autentikasi lima subyek. Pembelajaran mesin kemudian diintegrasikan dengan program antar-muka real-time, dan didapati persentase keberhasilan verifikasi sebesar 80%. ================== Current authentication systems generally have weaknesses, including card and key losses, and security vulnerabilities when entering passwords and Personal Identification Numbers. Biometric authentication systems, such as fingerprints and retinas, also have disadvantages, which can be imitated by fraudsters. In this research, we used the electroencephalography signal feature (EEG) as a new modality in biometric authentication, because it is unconscious and cannot be imitated with artifacts that are not living organisms. The electrode configuration is Fp2 for input signal, A1 for reference, and A2 for common-mode. The electrode is a disposable Ag/AgCl electrode. EEG signal instrumentation hardware consists of radio frequency interference filter, protection circuit, instrumentation amplifier, common-mode rejection circuit, DC differential circuit remover, non-inverting amplifier circuit, low pass filter with 72Hz cut-off frequency, high-pass filter with cut-off frequency of 0.23Hz, notch filter with 50Hz resonance frequency, isolation circuit, adder, and converted to digital with ADS1115. The total gain given by this circuit is 30.375,62x. Digital signals are then sent to Arduino Nano and processed using Personal Computer for processing EEG signals. The EEG signal is digitally filtered with a 3d Butterworth filter with a cut-off frequency of 4-14Hz, given a Hamming window, analyzed its frequency spectrum using Fast Fourier Transform, grouped in 128 bin, and normalized. Features are stored using Microsoft SQL Server and recognized by Artificial Neural Networks. The percentage of system verification success reached 96% for the authentication of five subjects. Machine learning was then integrated with real-time interface programs and found a verification success rate of 80%.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: biometrik, EEG, Jaringan Syaraf Tiruan, sinyal alfa, alpha signals, artificial neural networks, biometrics
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.B56 Biometric identification
Divisions: Faculty of Electrical Technology > Electrical Engineering > (S1) Undergraduate Theses
Depositing User: Jeff L Gaol
Date Deposited: 26 Oct 2018 08:54
Last Modified: 26 Oct 2018 08:54
URI: http://repository.its.ac.id/id/eprint/52845

Actions (login required)

View Item View Item