Elektrik ve Bilgisayar Mühendisliği Ana Bilim Dalı Tez Koleksiyonu

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/418

Browse

Now showing 1 - 4 of 4

Kağıt Tabanlı Magnetoforetik Sensör Geliştirilmesi
(Abdullah Gül Üniversitesi Fen Bilimleri Enstitüsü, 2022) FAROOQI, MUHAMMAD FUAD; Farooqi, Muhammad Fuad; İçöz, Kutay
One of the widely used type of biosensors are paper-based lateral flow systems. They are used to detect a wide variety of biomolecules like microorganisms, proteins, chemicals, oligonucleotides among many others. In this research, a setup was created using dual magnet sets in which the flow of cell sample on two kinds of different sample paper was explored. There were two factors which affected the movement of the sample the most, the magnetic field and the wetting. Images were obtained using a cell phone along and/or a bright field optical microscope and then analyzed using image processing. Images were also taken using scanning electron microscope. The effects of the wetting and the magnetic field were tested and studied. It was found that at least 90% of the cells were able to reach the edge of the paper. Although the cells were not able to maintain their shape on the paper due to the unideal conditions of the paper for cells but still this kind of paper-based lateral flow assay setup can be used for cells to see their behavior when they were labelled and exposed to a magnetic field. This research shows support that this technique can be used for separating cells as well as detecting different cells.
Akıllı Mikro-Şebekelerde Kontrol Stratejilerinin Geliştirilmesi
(Abdullah Gül Üniversitesi Fen Bilimleri Enstitüsü, 2021) Yoldaş, Yeliz; Yoldaş, Yeliz; Önen, Ahmet
This thesis concerns the transformation of aged power systems to modern power systems that include microgrids with renewable energy sources and energy storage systems. The integration of renewable energy sources brings excellent opportunities to provide better reliability and efficiency. However, the uncertainty and intermittent nature of renewable energy sources may potentially degrade the stability and quality of the electrical grid. Therefore, the aim of this dissertation is to maintain the supply-demand balance in microgrids while minimizing the cost in real time operation. A microgrid energy management system that can optimize the dispatch of the controllable distributed energy resources in grid-connected mode of a pilot microgrid on a university campus in Malta was developed to achieve this goal. Three different methods were used in this study: mixed integer linear programming (MILP), MILP based rolling horizon control and Q-learning, Designing intelligent method for the real-time energy management of the stochastic and dynamic microgrid is the primary goal of this research. Moreover, the detailed mathematical models of the network model and of the technical model are considered for the economic and environmental operation of the microgrid system to solve the optimization problem under more real-world conditions. The objective is to minimize the total daily operation costs, which include the degradation cost of batteries, the cost of energy bought from the main grid, the fuel cost of the diesel generator, and the emission cost. The optimization problem is modeled as a finite Markov decision process (MDP) by combining network and technical constraints, and a Q-learning algorithm is adopted to solve the sequential decision subproblems. The proposed algorithm decomposes a multi-stage mixed-integer nonlinear programming (MINLP) problem into a series of single-stage problems so that each subproblem can be solved using Bellman's equation. To prove the effectiveness of the proposed algorithm, three different case studies are taken into consideration. A predictive control framework is also proposed to provide optimal operation with minimum cost. This method allows the consideration of operational cost values, demand with uncertainty, generation units' profiles with uncertainty, and constraints related to the network model and technical model. The stochastic and deterministic cases are conducted to validate the efficiency of the approach.
Protein-Ligand Komplekslerinin Konvolüsyenel Sinir Ağları ile Moleküler Tanınması
(Abdullah Gül Üniversitesi Fen Bilimleri Enstitüsü, 2022) Güner, Hüseyin; Aydın, Zafer
As a sub-discipline of Artificial Intelligence, deep neural networks have received enormous interest in research and industrial applications over the last decades owing to their highly successful performance in addressing and solving broad areas of problems. Hence, especially hitherto achievements in computer-aided drug design brought an extra impetus with the novel deep learning approaches in structure-based drug design etiology. Our group offers a novel convolutional neural network model, deepMLR, that casts insight into the molecular recognition of ligand molecules and a receptor protein molecule. Having compared our model and a few other existing models with a case study of a traditional approach, herein, we present the success story of a deep learning model straight.
Hava-Yer Robot Ekipleri için Bağıl Lokalizasyon ve Koordinasyon
(Abdullah Gül Üniversitesi Fen Bilimleri Enstitüsü, 2021) Yıldırım, İsa Emre; Yıldırım, İsa Emre; Güler, Samet
Recently, autonomous robot teams have been implemented broadly in many social and military applications such as firefighting, agriculture, search and rescue, mapping, target tracking, and docking. A mix of different types of ground robots and aerial vehicles can be employed in a robot team to accomplish tasks efficiently and robustly. Such heterogeneous systems show unparalleled benefits in complex tasks compared to teams composed of identical robot types. In a heterogeneous robot team, precise relative localization, i.e., estimating a robot's position with respect to its neighbor robots, plays a key role. We develop a relative localization system for air-ground robot teams where an aerial vehicle and multiple ground robots work in coordination to perform a reliable relative position estimation. The aerial vehicle is employed to detect special patterns on the ground robots by an onboard monocular camera, while the ground robots perform relative position estimation based on inter-robot distances acquired by ultrawideband sensors and the bearing and heading angles received from the aerial vehicle by communication. Thus, the aerial vehicle serves as an absolute frame provider for the entire team. Notably, each robot in the team uses onboard communication and computation capabilities solely without any need for an external localization infrastructure, making the team realizable in all conditions including GNSS-denied environments. We propose a multi-rate extended Kalman filter algorithm to handle different data rates of the sensor measurements. We carried out an extensive simulation study with a drone and five ground robots in a leader-first follower formation. Simulation results showed a successful estimation performance with an error rate of up to five centimeters in the relative position estimations in both axes. Keywords: Relative localization, Heterogeneous multi-robot systems, estimation algorithms, ultrawideband sensors

Browse

Browsing Elektrik ve Bilgisayar Mühendisliği Ana Bilim Dalı Tez Koleksiyonu by Publisher "Abdullah Gül Üniversitesi Fen Bilimleri Enstitüsü"