Eeg sinyalleri ile cihaz kontrolü
Yükleniyor...
Dosyalar
Tarih
2017
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İnönü Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/embargoedAccess
Özet
Beyinde meydana gelen elektriksel aktivitelerin ölçülmesinde ve analiz edilmesinde sırasıyla Elektroansefalografi (EEG), Pozitron Emisyon Tomografisi (PET), Tek Foton Emisyonlu Bilgisayarlı Tomografi (TFEBT), Magnetoensefalografi (MEG) ve fonksiyonel Manyetik Rezonans Görüntüleme (fMRG) teknikleri kullanılmaktadır. Ancak maliyet, boyut ve vücuda zarar verme riski açısından en güvenilir beyin sinyali ölçme ve analiz tekniği EEG olarak kabul edilmiştir. EEG sinyalleri beyin aktivitesinin analiz edilmesinde kullanılan en temel yöntemlerden biridir. EEG beyinde meydana gelen elektriksel aktiviteyi değerlendirmek için bir test aracı olarak kullanılmaktadır. 1929'da Hans Berger tarafından ilk kez insan EEG'sinin tanıtılmasıyla beyin aktivitelerinin ölçümü yapılmıştır. Yüzyıla yakın bir süredir EEG tanı aracı olarak kullanılmaktadır. EEG Beyin-Bilgisayar Arayüzü (BBA) denilen sistemlerde de kullanılmıştır. Nörologlar, bir hastanın beyin dalgası aktivitesindeki şekillere tanı koyarak epilepsi, nöbet veya diğer nörolojik rahatsızlıklara neden olan anormalliklerin bulunmasını sağlamışlardır. BBA'nın gelişmesiyle kısmi engelliler ve kas sisteminde sağlık problemlerine maruz kalan bireylerin yaşam kaliteleri artırılmıştır. Bu bireylerin özellikle de hareket kabiliyetlerini yitirmiş felçli hastaların hayata tutunma olanakları artırılmış ve onlara yardımcı olacak sistemler geliştirilmiştir. Günümüzde birçok bilim insanı bu konu üzerinde çalışmaktadır. EEG tabanlı BBA'ya ait birçok yöntem literatürde mevcuttur. Bu yöntemler arasında gerçek zamanlı BBA'larda çoklu sınıf ve doğruluk oranı göz önüne alındığında en çok kullanılan görsel uyaran tabanlı yöntemlerdir. Bu tez çalışmasında Sabit Durum Görsel Uyaran Potansiyel (SDGUP) tabanlı EEG sinyal analizi gerçekleştirilmiş. Amaca uygun olarak bireyin ekran üzerinde belli frekanslarda titreşen sol (6,66 Hz), sağ (8.57 Hz) ve yukarı (12 Hz) olmak üzere üç farklı yönü gösteren şekillere bakması istenilmiştir. Birey ekran üzerindeki titreşen şekillere bakarken Emotiv Epoc+ cihazı ile EEG sinyalleri bilgisayar ortamına alınmış, OpenVibe ve Matlab yazılım platformunda bu sinyaller analiz edilip sinyal anlamlandırma işlemi gerçekleştirilmiştir. OpenVibe'da kişisel EEG sinyal verileri kullanılarak sistem ağı eğitilmiştir. Daha sonra eğitilen bu ağ kullanılarak gerçek zamanlı EEG sinyalleri ile sistem online olarak test edilmiştir. Matlab yazılım ortamına alınan EEG sinyalleri ise bant geçirgen filtreden geçirilmiş daha sonra Hilbert ve Kısa Zamanlı Fourier Dönüşüm işlemlerine tabi tutulmuştur. Deneğin baktığı farklı frekanstaki her şekil için görsel olarak uyarıldığı Oksipital bölgede SDGUP cevabı bulma işlemi yapılmıştır. SDGUP cevabı tespit edildikten sonra eğitim için Multimedia Authoring and Management using your Eyes and Mind (MAMEM) kuruluşu tarafından Emotiv Epoc+ cihazı ile on bir denekten toplanmış veri Yapay Sinir Ağları (YSA) ve Destek Vektör Makinaları (DVM) ile eğitim için kullanılmıştır. Eğitilen sınıflandırıcı ağını test için Emotiv Epoc+ cihazı ile kişisel EEG verileri kullanılmıştır. Bu şekilde bireyin ekran üzerinde hangi yöne baktığı beyin sinyalleri kullanılarak tespit edilmiş ve Lego Mindstorm EV3 robotu bu yönlere göre hareketi sağlanarak kontrol edilmiştir. Sonuçların başarılı bir şekilde test edilmesi gelecekte nano teknolojiler ve mobil cihazlar ile sistem bütünleştirilerek EEG tabanlı BBA çalışmalarında cihaz kontrolünün çok daha hassas, hızlı ve başarılı olacağı hakkında ümit vermektedir.
Electroencephalography (EEG), Positron Emission Tomography (PET), Single Photon Emission Computed Tomography (TPEBT), Magnetoencephalography (MEG) and Functional Magnetic Resonance Imaging (fMRI) techniques are used in the measurement and analysis of electrical activities in the brain. However, the most reliable brain signal measurement and analysis technique has been accepted as EEG in terms of cost, size and risk of damage to the body. EEG signals are one of the most basic methods used to analyze brain activity. EEG is used as a test tool to evaluate the electrical activity in the brain. In 1929, for the first time by Hans Berger, brain activities were measured by introducing human EEG. Nearly a century ago, EEG was used as a diagnostic tool. It was also used in systems called Brain-Computer Interface (BBA). Neurologists have been able to diagnose the shape of a patient's brain wave activity and have found abnormalities that lead to epilepsy, seizures or other neurological disorders. With the development of the BBA, the quality of life of individuals with partial disabilities and health problems in the musculoskeletal system has been increased. Patients with paralysis, especially those who have lost their mobility, have increased the possibilities of survival and have developed systems to assist them. Today, many scientists are working on this subject. Many methods of EEG-based BBA are available in the literature. These methods are the most commonly used visual stimulus based methods when multiple classes and accuracy rates are considered in real-time BBAs. In this thesis, Steady State Visually Evoked Potential(SSVEP) based EEG signal analysis was performed. In accordance with the aim, the individual is asked to look at three different directions on the screen, left (6.66 Hz), right (8.57 Hz) and up (12 Hz), which oscillate at certain frequencies. When looking at the flickering shapes on the screen, the EEG signals were received with the Emotiv EpoC + device and the signals were analyzed and signals were analyzed in OpenVibe and Matlab software platform. OpenVibe also trained the system network using personal EEG signaling. Then, this network was trained and the system was tested online with real-time EEG signals. The EEG signals received in the Matlab software platform were passed through a bandpass filter and then subjected to Hilbert and Short Time Fourier Transform operations. Visually stimulated in the occipital region the SSVEP response was made to find out for every shape that Subject look at in different frekans. After the SSVEP response was detected, data collected from eleven subjects by Emotiv Epoc+ device by Multimedia Authoring and Management using your Eyes and Mind (MAMEM) organization for training were used for training with Artificial Neural Networks (ANN) and Support Vector Machines (SVM). Emotiv Epoc+ device and personal EEG data were used to test the trained classifier network. In this way, the direction in which the individual looked on the screen was determined using brain signals, and the Lego Mindstorm EV3 robot was controlled according to these directions. Successful testing of the results is expected in the future to integrate the system with nanotechnologies and mobile devices, which will make the device control in EEG based BBA workings much more precise, faster and more successful.
Electroencephalography (EEG), Positron Emission Tomography (PET), Single Photon Emission Computed Tomography (TPEBT), Magnetoencephalography (MEG) and Functional Magnetic Resonance Imaging (fMRI) techniques are used in the measurement and analysis of electrical activities in the brain. However, the most reliable brain signal measurement and analysis technique has been accepted as EEG in terms of cost, size and risk of damage to the body. EEG signals are one of the most basic methods used to analyze brain activity. EEG is used as a test tool to evaluate the electrical activity in the brain. In 1929, for the first time by Hans Berger, brain activities were measured by introducing human EEG. Nearly a century ago, EEG was used as a diagnostic tool. It was also used in systems called Brain-Computer Interface (BBA). Neurologists have been able to diagnose the shape of a patient's brain wave activity and have found abnormalities that lead to epilepsy, seizures or other neurological disorders. With the development of the BBA, the quality of life of individuals with partial disabilities and health problems in the musculoskeletal system has been increased. Patients with paralysis, especially those who have lost their mobility, have increased the possibilities of survival and have developed systems to assist them. Today, many scientists are working on this subject. Many methods of EEG-based BBA are available in the literature. These methods are the most commonly used visual stimulus based methods when multiple classes and accuracy rates are considered in real-time BBAs. In this thesis, Steady State Visually Evoked Potential(SSVEP) based EEG signal analysis was performed. In accordance with the aim, the individual is asked to look at three different directions on the screen, left (6.66 Hz), right (8.57 Hz) and up (12 Hz), which oscillate at certain frequencies. When looking at the flickering shapes on the screen, the EEG signals were received with the Emotiv EpoC + device and the signals were analyzed and signals were analyzed in OpenVibe and Matlab software platform. OpenVibe also trained the system network using personal EEG signaling. Then, this network was trained and the system was tested online with real-time EEG signals. The EEG signals received in the Matlab software platform were passed through a bandpass filter and then subjected to Hilbert and Short Time Fourier Transform operations. Visually stimulated in the occipital region the SSVEP response was made to find out for every shape that Subject look at in different frekans. After the SSVEP response was detected, data collected from eleven subjects by Emotiv Epoc+ device by Multimedia Authoring and Management using your Eyes and Mind (MAMEM) organization for training were used for training with Artificial Neural Networks (ANN) and Support Vector Machines (SVM). Emotiv Epoc+ device and personal EEG data were used to test the trained classifier network. In this way, the direction in which the individual looked on the screen was determined using brain signals, and the Lego Mindstorm EV3 robot was controlled according to these directions. Successful testing of the results is expected in the future to integrate the system with nanotechnologies and mobile devices, which will make the device control in EEG based BBA workings much more precise, faster and more successful.
Açıklama
Anahtar Kelimeler
Bilgisayar mühendisliği bilimleri-bilgisayar ve kontrol
Kaynak
WoS Q Değeri
Scopus Q Değeri
Cilt
Sayı
Künye
Çiğ, H. (2017). Eeg sinyalleri ile cihaz kontrolü. İnönü Üniversitesi. Malatya.
