Yazar "Öniz, Yeşim" seçeneğine göre listele

Listeleniyor 1 - 3 / 3
  • Küçük Resim Yok
    Öğe
    Chaotic Dynamics and Analysis with Artificial Neural Networks of Aftershocks of 2019 Silivri Earthquake
    (Gazi University, 2024) Aydogmus, Fatma; Öniz, Yeşim; Simuratli, Eljan; Tosyalı, Eren; Kaplanvural, İsmail; Mutlu, Ahu Kömeç; Çaka, Deniz
    Earthquakes, whose physical, economic, psychological, and social damages can last for many years, are of vital importance for Türkiye, which is located in the most active earthquake zone that causes many earthquakes in the world. The North Anatolian Fault (NAF) is one of Türkiye's most important tectonic elements as it is the world’s fastest-moving right-lateral and strike-slip active fault zone consisting of many segments. The recent 5.8 magnitude 2019 Silivri earthquake, which occurred in the part of the NAF zone crossing the Marmara Sea, is an indicator that earthquake activity continues in the region. Aftershocks play a crucial role in seismicity research and seismic hazard assessments in terms of providing data and usable information in the examination of seismic dynamics with the changes observed in their time-dependent behavior and regional distribution. In this study, the aftershocks of the Silivri earthquake were examined as a natural laboratory using nonlinear analysis methods. Within the scope of the study, aftershocks of the Silivri earthquake were analyzed with a hybrid artificial neural network as well as different neural network structures, and for this purpose, data from 361 aftershocks with a magnitude greater than 1.5 in the year following the earthquake were used.
  • Küçük Resim
    Öğe
    Recognizing handwritten digits using spiking neural networks with learning algorithms based on sliding mode control theory
    (Scientific and Technological Research Council Turkey, 2023-10-26) Öniz, Yeşim; Ayyıldız, Mehmet
    In this paper, a spiking neural network (SNN) has been proposed for recognizing the digits written on the LCD screen of an experimental setup. The convergence of the learning algorithm has been ensured by using sliding mode control (SMC) theory and the Lyapunov stability method for the adaptation of the network parameters. The spike response model (SRM) has been utilized in the design of the SNN. The performance of the proposed learning scheme has been evaluated both on the experimental data and on the MNIST dataset. The simulated and experimental results of the SNN structure have been compared with the responses of a conventional neural network (ANN) for which the weight update rules have been also derived using SMC theory. The conducted simulations and experimental studies reveal that convergence can be ensured for the proposed learning scheme and the SNN yields higher recognition accuracy compared to a conventional ANN.
  • Küçük Resim Yok
    Öğe
    Trajectory Tracking of a Quadrotor Using Type-2 Neuro-Fuzzy Controllers
    (Gazi University, 2024) Öniz, Yeşim
    In this study, the trajectory tracking problem of a rotary wing unmanned aerial vehicle has been addressed by the use of type-2 neuro-fuzzy controllers. In order to determine the effectiveness of the developed control system, simulation and experimental studies have been performed for two different trajectories. The movement of the quadrotor in each direction has been controlled by a separate controller, and the difference between the actual and target positions for the relevant axis during the trajectory tracking along with the time derivative of this value has been fed to the controllers as the input signals. In order to better evaluate the results obtained, experimental and simulation studies for the same trajectories have been repeated with proportional-integral-derivative (PID) controllers and the responses of the controllers have been compared. Real-time experimental studies have been carried out indoors in a controlled environment with the Ar.Drone 2.0 produced by Parrot company. Especially the results recorded in the real-time experiments indicate that the proposed type-2 controllers with sliding mode control theory-based learning algorithm provide less steady-state error and more robust system response.|Bu çalışmada, tip-2 nöro-bulanık denetleyiciler kullanılarak bir döner kanatlı insansız hava aracının yörünge takibi gerçekleştirilmiştir. Geliştirilen kontrol sisteminin etkinliğini belirlemek amacıyla, oluşturulan iki farklı yörünge için benzetim ve deneysel çalışmalar yapılmıştır. Her bir eksen için farklı bir denetleyici tasarlanmış olup hava aracının yörünge takibi sırasında ilgili eksen için gerçek ve hedef konumları arasındaki fark ve bu değerin zamana göre türevi denetleyicilerin giriş sinyalleri olarak kullanılmıştır. Elde edilen sonuçları daha iyi değerlendirebilmek amacıyla aynı yörüngeler için deneysel ve benzetim çalışmaları orantılı-integral-türev (PID) denetleyici ile tekrarlanmış olup denetleyicilerin cevapları karşılaştırılmıştır. Gerçek zamanlı deneysel çalışmalar, Parrot firması tarafından üretilen Ar.Drone 2.0 ile iç mekanda kontrollü bir ortamda gerçekleştirilmiştir. Özellikle deneysel çalışmalardan elde edilen sonuçlar, tip-2 nöro-bulanık denetleyiciler için geliştirilen kayma kipli kontrol tabanlı öğrenme algoritmalarının daha az kalıcı hal hatası ve daha gürbüz sistem cevabı sağladığını göstermektedir.