Klasifikasi Sinyal EMG Pada Otot Tungkai Selama Berjalan Menggunakan Learning Vector Quantization - Classification Of EMG In Lower Limb Muscle During Walking Using Learning Vector Quantization

Putra, Darma Setiawan (2016) Klasifikasi Sinyal EMG Pada Otot Tungkai Selama Berjalan Menggunakan Learning Vector Quantization - Classification Of EMG In Lower Limb Muscle During Walking Using Learning Vector Quantization. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
2214206006-Master thesis.pdf - Published Version

Download (1MB) | Preview

Abstract

Sinyal electromyography (EMG) adalah aktifitas listrik yang terjadi pada lapisan otot selama adanya gerakan aktif. Gaya berjalan seseorang akan dipengaruhi oleh struktur tulang dan otot sehingga gaya berjalan tersebut adalah unik. Keunikan ini dapat digunakan untuk data biometrik. Dalam penelitian ini, kami akan melakukan klasifikasi data EMG untuk 8 otot tungkai selama berjalan yaitu Rectus Femoris, Vastus Medialis, Vastus Lateralis, Bicep Femoris, Semitendinosus, Gastrocnemius Medialis, Gastrocnemius Lateralis, dan Tibialis Anterior. 6 orang subyek sehat ditempelkan 8 elektroda EMG pada otot tungkai dan diminta untuk berjalan secara normal sesuai dengan kecepatan masing-masing di laboratorium gait. Masing-masing subyek berjalan sebanyak 1 siklus berjalan (gait cycle) dan 3 kali pengambilan data (walking trial). Total data pengambilan adalah sebanyak 18 buah dataset untuk analisis dan klasifikasi. Metode graph feature extraction dan principal component analysis digunakan untuk mengekstraksi fitur data EMG dari 8 otot tungkai selama berjalan. Metode learning vector quantization (LVQ) digunakan untuk mengklasifikasi data EMG berdasarkan subyek. Metode pembelajaran dan pengujian data pada jaringan LVQ menggunakan metode validasi silang (cross validation). Akurasi klasifikasi rata-rata menggunakan metode graph feature extraction diperoleh sebesar 88.89% dan metode PCA diperoleh sebesar 66.67%. Dari hasil ini menunjukkan bahwa sinyal EMG selama berjalan dari 8 otot tungkai dapat digunakan sebagai identitas biometrik gait. ======================================================================================================================== Electromyography (EMG) signal is an electrical activity that occurs in the muscle layer during active motion. The way people walking is defined by the structure of individual muscle and bones so that the way of walking is unique and must be able to used in biometric data. In this study, we classified the EMG data dari 8 lower limb muscle during normal walking test (Rectus Femoris, Vastus Medialis, Vastus Lateralis, Bicep Femoris, Semitendinosus, Gastrocnemius Medialis, Gastrocnemius Lateralis, and Tibialis Anterior). Six healthy volunteer were involving in this study by walking in GaitLab with 8 EMG electrodes attached on their muscle. Each volunteer performed one gait cycle and 3 walking trial. So in total 18 EMG dataset were analized for classification. Graph feature extraction and principal component analysis method was used to extract the feature of EMG data of all 8 muscle during walking. Learning Vector Quantization (LVQ) was used to classify the EMG data based on subject. Training and testing method in LVQ network used cross validation (CV). The average accuracy of classification using graph feature extraction method is 88.89% and using PCA method is 66.67%. In the result show that EMG data during walking of 8 lower limb muscles can be used to identity of gait biometric.

Item Type: Thesis (Masters)
Additional Information: RTE 616.740 754 7 Put k
Uncontrolled Keywords: EMG, Otot Tungkai, Biometrik Gait, Principal Component Analysis, Analisis Gait.
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3226 Transients (Electricity). Electric power systems. Harmonics (Electric waves).
Divisions: Faculty of Industrial Technology > Electrical Engineering
Depositing User: ansi aflacha
Date Deposited: 17 Dec 2019 06:42
Last Modified: 17 Dec 2019 06:42
URI: http://repository.its.ac.id/id/eprint/72398

Actions (login required)

View Item View Item