Hakan Terzioğlu

This study comprehensively addresses the design and modeling processes of autonomous underwater vehicles (AUVs). The research examines the fundamental design principles that affect underwater vehicle performance, focusing on factors such as hydrodynamic efficiency, structural durability, buoyancy control, and stability. Considering the harsh marine environment, the selection of materials was thoroughly discussed based on criteria such as mechanical properties, corrosion resistance, and weight optimization. The production stages, from 3D modeling to manufacturing, were explained step by step. Engineering challenges encountered during the design process—such as waterproofing, stability, and balance—were emphasized, and solution approaches to these problems were presented. Computational Fluid Dynamics (CFD) analyses were employed to evaluate the hydrodynamic performance of the vehicle and to optimize the body geometry. Finally, based on the analysis results and design iterations, solution strategies and development suggestions were proposed, aiming to guide future studies in the development of autonomous underwater vehicles.

In recent years, significant advancements in artificial intelligence, particularly in the field of deep learning, have increasingly been integrated into agricultural applications, including critical processes such as disease detection. Tomato, being one of the most widely consumed agricultural products globally and highly susceptible to a variety of fungal, bacterial, and viral pathogens, remains a prominent focus in disease detection research. In this study, we propose a deep learning-based approach for the detection of tomato diseases, a critical challenge in agriculture due to the crop’s vulnerability to fungal, bacterial, and viral pathogens. We constructed an original dataset comprising 6414 images captured under real production conditions, categorized into three image types: leaves, green tomatoes, and red tomatoes. The dataset includes five classes: healthy samples, late blight, early blight, gray mold, and bacterial cancer. Twenty-one deep learning models were evaluated, and the top five performers (EfficientNet-b0, NasNet-Large, ResNet-50, DenseNet-201, and Places365-GoogLeNet) were selected for feature extraction. From each model, 1000 deep features were extracted, and feature selection was conducted using MRMR, Chi-Square (Chi²), and ReliefF methods. The top 100 features from each selection technique were then used for reclassification with traditional machine learning classifiers under five-fold cross-validation. The highest test accuracy of 92.0% was achieved with EfficientNet-b0 features, Chi² selection, and the Fine KNN classifier. EfficientNet-b0 consistently outperformed other models, while the combination of NasNet-Large and Wide Neural Network yielded the lowest performance. These results demonstrate the effectiveness of combining deep learning-based feature extraction with traditional classifiers and feature selection techniques for robust detection of tomato diseases in real-world agricultural environments.

In recent years, significant advancements in artificial intelligence, particularly in the field of deep learning, have increasingly been integrated into agricultural applications, including critical pro-cesses such as disease detection. Tomato, being one of the most widely consumed agricultural products globally and highly susceptible to a variety of fungal, bacterial, and viral pathogens, remains a prominent focus in disease detection research. In the present study, a five-class classi-fication was conducted to distinguish between healthy tomatoes and those affected by the most common diseases: tomato late blight, early blight, gray mold, and bacterial cancer. The dataset used in this study is entirely original and consists of images captured in the production envi-ronment, covering three different categories: leaves, red tomatoes, and green tomatoes. Through twofold data augmentation, the dataset was expanded to include a total of 6,414 samples. The data was split into training and testing sets with a ratio of 80% to 20%, respectively, and 21 deep learning models were evaluated. Among these, the five best-performing models (NasNet-Large, ResNet-50, DenseNet-201, EfficientNet-b0, and Places365-GoogLeNet) were selected, and 1,000 features were extracted from each. From these extracted features, the top 100 most discriminative features were selected using three different feature selection algorithms: Minimum Redundancy Maximum Relevance (MRMR), Chi-Square (Chi²), and ReliefF. These features were subsequently used for reclassification with machine learning algorithms, employing five-fold cross-validation while maintaining the same train-test split. The highest test accuracy of 92.0% was achieved using the EfficientNet-b0 model combined with Chi² feature selection and the Fine K-Nearest Neighbors (KNN) classifier. Overall, EfficientNet-b0 emerged as the most effective deep learning model, while the ReliefF feature selection method and the Subspace KNN classifier yielded higher accu-racies compared to other methods. In contrast, the combination of NasNet-Large and the Wide Neural Network classifier demonstrated the weakest performance with the lowest test accuracy. These findings indicate that features extracted via deep learning, when combined with appropriate feature selection methods and traditional classifiers, can be highly effective in the detection of tomato diseases.

Today, energy is derived from a very large percentage of fossil-based sources, such as 80-85% worldwide. However, it is not sustainable that the use of fossil-based resources pollutes the environment, causes climate change, runs out of reserves and lacks price stability. The most important measure against climate change that poses a threat to the world is to reduce the share of fossil-based resources in energy supply.In this context, in order to encourage universities in our country to design and manufacture electric vehicles and equipment, TUBITAK organizes competitions for universities and high schools under the name electromobile every year. In this study, TUBITAK Electromobile Battery was designed and manufactured by designing a unique electric vehicle to compete in the Electric Powered Vehicle category. After the mechanical designs of this vehicle we produced were made, the chassis and shell production was carried out by analyzing. After the mechanical designs of this vehicle we produced were made in the Solidworks program, the ANSYS 2019 R3 program was used for the analysis, and the chassis and shell were produced.

The use of unmanned aerial vehicles (UAV) is increasing today. UAVs can be divided into two parts, which are remote controlled and can travel automatically due to a certain battery problem. Recent research has also focused on the development and application of new algorithms to autonomously control these vehicles and determine the shortest flight paths. Together with these researches, UAVs are used in many civil activities such as weather forecasts, environmental studies and traffic control. Three-dimensional (3D) path planning is an important issue for autonomously moving UAVs. The shortest path for Unmanned Aerial Vehicles (UAV) is determined by using two-dimensional (2D) path planning algorithms using the obstacles in the environment, and allows UAVs to perform their environmental tasks as soon as possible. The purpose of this study is to determine the shortest path to the target point and avoiding obstacles for UAVs using the Dijkstra algorithm. It was developed to evaluate the arrival time of the UAVs in the path planning algorithm with the simulation performed in the MATLAB program. In this study, the obstacles were defined for the purpose of the building with different heights and different widths and 2D and 3D models were carried out, assuming that the UAV flies at certain heights. In addition, the flight of the UAVs in the route planning determined in the real applications was carried out and the data such as battery consumption, amount of battery spent, speed, amount of travel were examined.

Researchers and industry are seeking to manage energy better by increasing the efficiency of the energy system because of the increasing importance of energy in recent years. For this reason, a shift towards alternative energy sources to provide electricity has begun. As for alternative energy sources, interest in renewable energy sources is increasing day by day. Solar, wind and thermal sources are the most prominent renewable energy sources. Today we use of thermal sources in greenhouse, fish breeding, thermal facilities, city heating and electricity production. Nowadays, it is known that when we move thermal water from the source to the place to be used, it can be converted into electrical energy by utilizing the heat it emits around it. In this study, the effects of the type of the material and the placement of the thermoelectric generator (TEG) on the efficiency of two Thermoelectric Modules (TEM) were investigated by using thermal sources from renewable energy sources. Hot and cold water was passed through the sandwich blocks which were made from Copper, Aluminum and Brass materials, and the power produced in TEG was analyzed. As TEG, TEC1-12706 and TEC1-12710 materials were used. As a result of experimental studies, it is observed that the type of the material that conveys thermal water and the order of transmission of hot/cold water through sandwich blocks have an effect on the power produced by both TEC1-12706 and TEC1-12710

Transportation vehicles refer to the vehicles used to transport people's needs and goods from one area to another. Intelligent traffic signalling systems are needed due to the increase in the number of vehicles in traffic, the roads used, the insufficiency of the surrounding areas and the inability to widen these roads or construct upper / lower roads. Nowadays, there are situations where traffic control is insufficient in current density due to the traffic system mechanisms being dependent on fixed time. Therefore, some smart methods have been developed to prevent waste of time and economic problems. In this study, application of Petri Nets (PA) based on real data collected at certain times of the day for intelligent control of traffic lights at a four-way intersection. PA's performances have been compared with their classical (fixed time) performances. When the results obtained were evaluated in terms of both the number of passing vehicles and time, it was seen that the PA passed 849 vehicles in 755 seconds, 341 vehicles in 920 seconds in the classical method and a better result was obtained with PA.

Spesifik enerji değeri mekanik kayaç kazısında kayaçların kesilebilirlik özelliklerini belirlemek içinyaygın olarak kullanılmaktadır. Mekanik kayaç kazısı işlemlerinde mühendisler, spesifik enerjideğerine bağlı olarak makine performansını tahmin etmek için kolay uygulanabilir, daha ekonomikve basit örnek hazırlama yöntemlerinin kullanıldığı yöntemlere ihtiyaç duyarlar. Bu çalışmada,spesifik enerjinin tahmini için pratik, basit ve ucuz bir test olduğu düşünülen kayaçların P-dalgahızı (Vp) ve Schmidt çekici sertlik (RL) değerleri değişken olarak önerilmiştir. Bu amaçla 24 farklıkaya numunesi üzerinde kaya kesme ile Vp ve R L testleri yapılmıştır. Elde edilen Vp ve R L ilespesifik enerji değerleri SPSS 15.0 programı kullanılarak basit ve çoklu regresyon analizi iledeğerlendirilmiştir. Bu değerlendirme sonucunda kayaçların spesifik enerji, Vp ve R L değerleriarasında güçlü bir ilişki olduğu belirlenmiştir. İstatistiksel analizlere göre, laboratuar çalışmalarınabağlı olarak kayaların Vp ve RL değerleri kullanılarak spesifik enerji değerleri güvenilir bir şekildetahmin edilebilir.

The Minimization of Torque Ripples of Segmental Type Switched Reluctance Motor by Particle Swarm Optimization

The purpose of this experimental study was to measure stresses both on the pterygoid plates and the skull base following two different pterygomaxillary approaches in Le Fort I osteotomy. The prepared skull models were randomly divided into 2 groups of 7. In the first group (A), the pterygomaxillary area was left intact. In the second group (B), pterygomaxillary separation was performed with a fine bur. The stresses were measured by using strain gauges. These strain gauges were attached to 6 different anatomical sites. The skull models were mounted on a servo-hydraulic testing unit. Each model was then subjected to a continuous linear tension until a plastic deformation was seen. The statistical analyses showed that there were no significant differences (p >.05) between the 2 groups regarding the strain values. Moreover, no statistical differences (p >.05) were found between the two groups in terms of maximum applied forces. Considering the clinical conditions, the present study shows that when Le Fort I osteotomy performed without pterygomaxillary separation, there is no significant stress on the skull base during the downfracture. Moreover, it is considered that there is no need for an excessive force applied to perform downfracture in Le Fort osteotomies without pterygomaxillary separation