Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Merdan Özkahraman has a total of 25 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • DESIGN AND CONTROL OF A SIX-LEGGED SPIDER ROBOT FOR ENHANCED MOBILITY AND ADAPTABILITY

    by Yusuf Bolat, Hüseyin Ömerbaşoğlu, Merdan Özkahraman, Hilmi Cenk Bayrakçı

    Spider robots are widely preferred due to their superior mobility enabled by multi-legged structures and articulated joints, which allow effective locomotion on uneven terrain and obstacle negotiation. However, the mechanical complexity of legged systems may lead to limited payload capacity, a high number of moving parts, and increased sensitivity to impacts, while balance control becomes more challenging due to variations in the center of gravity. This study aims to mitigate these drawbacks by proposing the mechanical design and prototype implementation of a six-legged spider robot, supported by material selection and an embedded electronic setup for actuation and sensing. Forward and inverse kinematic models of a 3-DOF leg are derived using the Denavit–Hartenberg convention to enable consistent foot placement and trajectory generation. In addition, predefined gait patterns are implemented and evaluated on the prototype to assess feasible locomotion behavior, and a structural deformation analysis of critical leg segments is performed to investigate mechanical robustness under loading. The presented platform provides a practical basis for further developments in hexapod locomotion for field-oriented applications such as search and rescue, reconnaissance, and hazardous environment exploration.

  • ENDÜSTRİYEL EKİPMANLARDA YAPAY ZEKA DESTEKLİ ÖN KESTİRİMCİ BAKIM STRATEJİLERİNİN GELİŞTİRİLMESİ VE UYGULANMASI

    by Yakup Bedirhan Menteşe, Merdan Özkahraman

    Endüstriyel üretim süreçlerinde makinelerin güvenilirliği ve kullanılabilirliği, üretim verimliliği açısından kritik öneme sahiptir. Geleneksel bakım yaklaşımları, arıza gerçekleştikten sonra veya belirli periyotlarla müdahaleyi esas alırken, bu yöntemler gereksiz bakım maliyetlerine ve plansız duruşlara neden olabilmektedir. Bu çalışmanın amacı, yapay zeka tabanlı ön kestirimci bakım yaklaşımı geliştirerek, ekipman arızalarının gerçekleşmeden önce tahmin edilmesini sağlamaktır. Araştırmada, CMX 1100 V model CNC freze tezgahından elde edilen açık kaynaklı sensör verileri kullanılmış; veriler Python ortamında işlenmiş ve XGBoost algoritması ile analiz edilmiştir. Model, işleme durumunu yüksek doğruluk oranı ile sınıflandırmış ve olasılık tabanlı bakım kararı verebilen bir karar destek sistemi olarak tasarlanmıştır. Elde edilen sonuçlar, yapay zeka destekli ön kestirimci bakımın üretim süreçlerinde bakım planlamasını optimize ederek maliyetleri azaltabileceğini ve üretim sürekliliğini artırabileceğini göstermektedir. Çalışma, makine öğrenimi tabanlı karar mekanizmalarını endüstriyel bakım stratejilerine entegre ederek Endüstri 4.0'ın dijital dönüşümüne katkıda bulunmaktadır.

  • DEEP LEARNING-BASED VEHICLE BRAND AND MODEL RECOGNITION FOR MARKET ANALYSIS USING YOLOV8 AND YOLOV11 ON REAL-WORLD TRAFFIC DATA

    by Recep Arslan, Merdan Özkahraman, Hilmi Cenk Bayrakçı

    The ability to analyze vehicle brand distribution from real-world traffic data has significant implications for market research and trend analysis. This study presents a deep learning-based approach for automated vehicle brand and model recognition using the YOLOv8 and YOLOv11 object detection architectures. The proposed system was trained and evaluated on a dataset of 5,213 labeled vehicle images, collected from actual road traffic footage to ensure real-world applicability. Performance assessments were conducted using mean Average Precision (mAP), precision, recall, and inference speed, demonstrating that YOLOv11 outperforms YOLOv8 in precision and object localization accuracy, achieving a mAP of 88.6%. However, YOLOv8 achieved a higher recall, detecting a broader range of vehicles, albeit with an increased false positive rate. In addition to model performance analysis, this study examines the distribution of different automobile brands in traffic, providing valuable insights for data-driven market analysis. The findings highlight the potential of deep learning-based vehicle detection as a tool for understanding brand presence and consumer trends in the automotive market.

  • Real‐Time Detection of Undesired Human Interventions in Robotic Work Cells Using a Convolutional Neural Network‐Based Novel Architecture and Reliability Analysis With Explainable Artificial Intelligence

    by Merdan Ozkahraman

    This study presents a task‐specific deep‐learning framework for real‐time detection of undesired human interventions in robotic work cells, based on a customized convolutional neural network (CNN) architecture termed MerdanNet. Unlike general‐purpose lightweight models, MerdanNet integrates progressive dropout, systematic batch normalization, and compact hierarchical depth to ensure reliable performance in low‐data, safety‐critical environments. The data set comprises 588 original images of human hand, upper body, and foot categories, expanded to 2940 samples through targeted augmentation (mirroring, grayscale transformation, and controlled noise). Three models, MerdanNet, YOLOv8, and MobileNet, were trained and evaluated using accuracy, precision, recall, and F 1‐score metrics; MerdanNet achieved the highest performance, with an accuracy of 98.76%. Beyond static evaluation, the framework was validated in a closed‐loop robotic setup where detections directly triggered safety actions. Experiments across 90 trials confirmed consistent activation of slowdown and emergency stop functions with an average end‐to‐end latency of 92 ms (95th percentile: 124 ms), well within industrial safety thresholds. Interpretability was assessed using Grad‐CAM and LIME, which revealed meaningful attention patterns and provided diagnostic insights, though quantitative explainable artificial intelligence (XAI) evaluation remains a target for future work. While the data set is limited in diversity, the study highlights this as a current limitation and outlines future directions, including expanded data collection, synthetic stress‐test data sets, bounding box annotations for detector benchmarking, and transfer learning approaches. Overall, the findings demonstrate that combining tailored CNN architectures with XAI and closed‐loop validation can yield deployable, transparent, and robust safety modules for industrial robotic environments.

  • Open Access

    Experimental modal analysis of optimized additively manufactured robotic arms coated with ceramic powders and resin

    by Mamoun Alshihabi, Merdan Özkahraman, Mevlüt Yunus Kayacan

    This study investigates the vibration damping properties of additively manufactured robotic arm components enhanced with optimized infill patterns and resin coatings doped with alumina and silicon carbide. Seven designs, including Voronoi, triangular, and square honeycomb structures at 1 and 2 mm scales, were tested. Modal analysis was performed to evaluate natural frequencies, damping ratios, and damping coefficients. Sample E demonstrated superior damping performance, with the highest damping coefficient (3.2 Ns/m) and ratio (0.18). Sample A represents the unoptimized solid baseline part, while Sample E corresponds to the 2 mm triangular honeycomb design. In contrast, Sample A exhibited the lowest values, indicating inferior vibration control. The Voronoi patterns showed the most significant weight increase post-coating due to their surface area, while triangular honeycomb designs balanced weight and damping performance. These findings highlight the effectiveness of combining advanced materials and designs to enhance robotic arm stability, precision, and longevity. The study provides valuable insights for optimizing robotic components for industrial applications requiring high performance and reliability.


  • Open Access

    A NOVEL HYBRID ENCRYPTION SCHEME BASED ON MODERN CRYPTOGRAPHIC COMPONENTS FOR EMBEDDED SYSTEMS

    by Mustafa Emre Erbil, Hilmi Cenk Bayrakçı, Merdan Özkahraman, Гільмі Дженк Байракджи

    Topicality. Embedded systems are gaining increasing importance in various domains such as automotive, industrial automation, healthcare technologies, and the Internet of Things (IoT). The main characteristics of these systems include limited memory, moderate processing power, and strict energy efficiency requirements. These constraints make secure data transmission essential, particularly in both wired and wireless communication processes. Potential data leaks or unauthorized access pose major threats to the reliability and safety of embedded systems. The subject of the research is methods and mechanisms of secure data transmission in embedded systems based on the proposed hybrid cryptographic architecture. The main objective of this study is to propose a reliable and security-oriented hybrid cryptographic solution that addresses the need for confidential and integrity-protected data transmission in such devices. Results. In the proposed method, an asymmetric mechanism based on X25519 is employed for secure key exchange. The BLAKE3 function is utilized for key derivation due to its deterministic structure and cryptographic strength, while Ascon-128a is adopted in the symmetric encryption layer for its lightweight yet highly secure AEAD design. This combination integrates the robust key exchange capability of X25519, the collision-resistant and high-entropy key derivation of BLAKE3, and the AEAD (Authenticated Encryption with Associated Data)-based integrity and confidentiality protection of Ascon-128a within a unified framework. The feasibility of the proposed hybrid structure was tested on a Raspberry Pi 5 using the Python programming language. In the experimental setup, the method was validated under different communication scenarios between an embedded sensor and a controller. Conclusion. The findings demonstrate that the developed hybrid architecture provides a highly secure, integrity-preserving, and forward-secrecy-compliant alternative for data protection in embedded systems, offering stronger cryptographic guarantees compared to conventional hybrid encryption schemes in the literature.

  • Open Access

    THE USE OF FULLY AUTONOMOUS CALF FEEDING ROBOTS TODAY AND ANALYZING THE CONSUMPTION HABITS OF CALVES

    by Ali Bilir, Ferdi Alakus, Ayşegül Kiliçaslan, Merdan Özkahraman

    In order for calves to complete their development in a healthy way, their diet for the first four months is of critical importance. In this period, colostrum intake immediately after birth, feeding with milk or milk derivative (nowadays many farms prefer powdered formula) liquid feeds in the first two months, and starting from the second month, calves should be directed to solid foods. Rapid developments in agricultural technologies have led to significant changes in the livestock sector. In this context, the use of fully autonomous calf feeding robots is becoming more and more widespread in order to increase productivity in animal husbandry and to support the healthy development of calves. Autonomous feeding robots optimize the feeding process of calves, reducing labor costs and providing a more controlled and balanced nutrient intake. These technological advances increase sustainability and competitiveness in animal husbandry while prioritizing animal welfare. In this study, it was aimed to investigate the extent to which farms comply with the standards specified in the literature when feeding calves and to investigate the root causes of health problems encountered by monitoring the behavior of calves during the free feeding process. It was aimed to analyze the relationship between the change in routine movements of calves and health problems by recording the first four months of calves through autonomous calf feeding robots. In this context, through data collection and analysis, the effectiveness of calf feeding processes and the effects of these processes on calf health were evaluated.

  • Open Access

    ROBOTİK KOL ENTEGRELİ MOBİL ARAMA VE KURTARMA ROBOTUNUN TASARIMI VE UYGULAMASI

    by Cemal Beık, Merdan Özkahraman

    Son yıllarda yaşanan doğal afetler, arama ve kurtarma operasyonlarının hızını ve etkinliğini artıracak teknolojik çözümlere duyulan ihtiyacı ortaya koymuştur. Bu çalışmada, doğal afetler sırasında kullanılmak üzere robotik kol entegreli mobil bir arama ve kurtarma robotu tasarımı ve gerçekleştirilmesi ele alınmaktadır. Robot, dar alanlarda manevra yapabilme ve enkaz altındaki hayatta kalanlara temel ihtiyaç malzemelerini ulaştırabilme kapasitesine sahiptir. Robotun termal kameralar, ses sensörleri gibi ileri teknolojilerle donatılması ve insan kurtarma ekipleri ile etkili bir şekilde iletişim kurabilmesi, kurtarma operasyonlarının verimliliğini artırmaktadır. Robotun kabiliyetleri, zamanın kritik olduğu acil durumlarda hızlı müdahale imkânı sunarak hayatta kalanların bulunmasını ve kurtarılmasını hızlandırmaktadır. Ayrıca, robotun riskli bölgelere insanların girmesine gerek kalmadan müdahale edebilme yeteneği, kurtarma ekiplerinin güvenliğini artırmaktadır. Robotun tasarımı ve gerçekleştirilmesi sürecinde kullanılan ileri teknolojiler ve mühendislik çözümleri, bu tür robotların gelecekteki arama ve kurtarma operasyonlarında yaygın olarak kullanılmasını sağlayabilir. Bu çalışma, ülkenin afetlerle başa çıkma yeteneğini artırmak için teknolojik yenilik ve istihdam fırsatları sunmaktadır.

  • Open Access

    Comprehensive Dynamic Analysis of a 6-DOF Robotic Arm Using Advanced Simulation Techniques

    by Merdan Özkahraman

    This paper presents a detailed dynamic analysis of a six degrees of freedom (6-DOF) robotic arm using advanced simulation techniques in MATLAB Simulink. The study focuses on evaluating the robotic arm's performance under various operational conditions, including joint movements, trajectory tracking, payload handling, disturbance rejection, and navigation in dynamic environments. Kinematic and dynamic models were developed using the Denavit-Hartenberg (D-H) parameters and Lagrange equations, enabling a comprehensive understanding of the arm's motion and force interactions. Sinusoidal inputs with varying amplitudes were applied to the joints, producing detailed profiles for angular displacement, velocity, acceleration, and torque. The findings reveal that the base joint experiences the largest torques due to its role in broader rotational movements, while wrist joints exhibit smaller, more frequent adjustments required for precise control. The study emphasizes the importance of tailored control strategies and actuator designs to meet the unique demands of each joint. These insights contribute to the development of more efficient and robust robotic systems, with applications in manufacturing, medical procedures, and space exploration.