Bu sayfa, EE690 Lisansüstü Seminerleri dersi kapsamında gerçekleştirilecek seminerler hakkında ilgili katılımcıları bilgilendirmek amacıyla oluşturulmuştur. Program ve içerik dönem boyunca güncellenecektir.
2025-2026 Güz Dönemi Lisansüstü Seminerleri
Başlık: Bio-inspired Observability Tools for Active Sensing and Sensor Selection
Özet: Flying insects have an unexpectedly remarkable sensing capacity. For instance, Drosophila, the fruit fly, can track chemical plumes by estimating the ambient direction without directly measuring it. In this talk, I will introduce the observability tools we developed to better understand the active sensing decisions made by insects and the importance of sensor selections. Specifically, I will address how to determine system observability levels (i.e., estimator performance limits) in the presence of process and/or measurement noise, and for particular state variables, e.g., state of charge or state of health of a battery. I will conclude the talk with potential engineering applications.
Özgeçmiş: Dr. Burak Boyacıoğlu earned his Ph.D. from the University of Washington's (UW) William E. Boeing Department of Aeronautics and Astronautics in 2022. He holds a B.S. in Aeronautical Engineering, a minor in Mechatronics, and a Master's in System Dynamics and Control, all from Istanbul Technical University (İTÜ). His research interests include optimal sensor placement and active sensing strategies for highly sensed systems, with applications in biological flight and engineered flight, and space systems. Most recently, he was a postdoctoral scholar at the University of Nevada, Reno (UNR), and taught a grad-level course for professional students at UW and an undergrad course at İTÜ. He is currently teaching controls courses in the Department of Aerospace Engineering at METU as a visiting scholar.
Başlık: Cognitive Electronic Warfare (CEW)
Özet: Electronic Warfare (EW) involves the use of the electromagnetic spectrum to gain a tactical advantage over opponents. It has two main components: Electronic Support (ES), which improves situational awareness by detecting, intercepting, identifying, and locating signals, and Electronic Attack (EA), which disrupts enemy radars through jamming and deception. Traditional ES methods, which rely on static emitter libraries and deterministic signal processing, along with classical EA techniques based on pre-programmed jamming strategies, were effective against earlier and predictable threats. However, modern operational environments present challenges such as the rapid emergence of adaptive radar systems, low-probability-of-intercept (LPI) waveforms, dense multi-emitter scenarios, which undermine the static nature of conventional EW approaches. This has led to the development of cognitive techniques that introduce adaptability and intelligence into electronic warfare. Cognitive Electronic Support Measures (CESM) utilize machine learning, adaptive signal processing, and real-time decision-making to identify emissions and dynamically adjust sensing strategies. Similarly, cognitive electronic attack (CEA) extends traditional EA by employing adaptive and autonomous jamming or deception techniques to deny detection and tracking by modern radars. This presentation introduces the fundamental concepts of electronic warfare and cognitive electronic warfare, highlighting the motivations for incorporating cognition into modern EW systems and radar technologies, with a particular emphasis on the applications of Cognitive Electronic Support Measures (CESM).
Başlık: Doğumda Yapay Zekâ – Yapay Plasenta
Özet: Sunum üç bölümden oluşmaktadır.
- Yapay Zekâ ile Doğum Yönetimi: Dünya Sağlık Örgütü (DSÖ) bünyesinde tamamlanan ve hâlihazırda sağlık hizmetlerine ve uzman hekimlere erişimde zorluk yaşayan Afrika ülkelerinde kullanılmaya başlanan bir yapay zekâ projesi tanıtılacaktır. Bu sistem, doğumun takibini ve yönetimini kolaylaştırmakta, doğuma ilişkin tüm risklerin en aza indirilmesini hedeflemektedir.
- Yapay Plasenta: Tüm dünyada yenidoğan ölümleri ve kalıcı sekellerin yaklaşık %65’inden tek başına sorumlu olan erken doğumun önlenmesine odaklanan “yapay plasenta” projesi ele alınacaktır. Bu proje, gebelik risklerinin azaltılmasında ve erken doğumun önlenmesinde önemli bir bilimsel dönüm noktası olma potansiyeline sahiptir.
- Bilkent Cyberpark Projeleri: Son olarak Prof. Dr. Çağrı Gülümser ve ekibinin Bilkent Cyberpark’ta yer alan şirketlerinde yürüttükleri, embriyodan üniversiteye kadar uzanan sağlık, gelişim ve eğitim alanındaki projelerden bahsedilecektir. Ayrıca bu projelere katılmak veya katkı sağlamak isteyen katılımcılarla birebir görüşme olanağı olacaktır.
Özgeçmiş: Prof. Dr. Çağrı Gülümser, Ondokuz Mayıs Üniversitesi Tıp Fakültesi’nden 2000 yılında mezun olmuş, Ankara Atatürk Eğitim ve Araştırma Hastanesi’nde Kadın Hastalıkları ve Doğum ihtisasını tamamlamıştır. İngiltere’de King’s College Hospital’da Maternal ve Fetal Tıp, Fetal Ekokardiyografi ve University College London’da Üreme Tıbbı alanlarında ileri eğitimler almıştır. 2008’de ESGE tarafından “En İyi Bilimsel Çalışma”, 2009’da RCOG tarafından “En İyi Bilimsel Sunum”, 2011’de ise “En Yenilikçi Bilimsel Çalışma” ödüllerine layık görülmüştür. 2012–2019 yılları arasında Başkent ve Sağlık Bilimleri Üniversitelerinde öğretim üyesi olarak çalışmış, 2021’de Yüksek İhtisas Üniversitesi Tıp Fakültesi Kadın Hastalıkları ve Doğum Anabilim Dalı’nda Profesör olarak atanmıştır. Dünya Sağlık Örgütü, T.C. Sağlık Bakanlığı, Dünya Bankası ve Bill & Melinda Gates Vakfı ile birçok uluslararası projede bilim insanı olarak yer almıştır. Ayrıca BJOG, BMC Pregnancy & Childbirth ve Turk J Obstet Gynecol gibi dergilerde editörlük ve hakemlik yapmıştır. Prof. Gülümser, Gülümser Akademi ve Bilkent Cyberpark’ta faaliyet gösteren Gülümser A.Ş. ARGE Merkezi’nin kurucu ortağı, 2024’te kurulan Gülümser Vakfı’nın Kurucu Başkanı ve Bilimsel Danışma Kurulu Başkanıdır. Çalışmaları kadın sağlığı, epidemiyoloji ve perinatal yapay zekâ uygulamaları üzerinde yoğunlaşmaktadır.
Başlık: Academia–Industry Integration in Robotic Systems Engineering: Experience, Learning, and Transformation
Özet: TBA
Özgeçmiş: TBA
Başlık: Bridging Chemistry and Functionality in Soft Matter Systems
Özet: TBA
Özgeçmiş: TBA
Başlık: From Graded Composites to Smart Materials: Adaptive Surfaces for Soft Robotics
Özet: A materials-driven approach for achieving adaptive behavior in soft systems is presented, beginning with the mechanical design of graded and layered polymer nanocomposites. By systematically tuning layer architecture, concentration gradients, and coating thickness, smooth stress distribution, enhanced strength–toughness balance, and force-dependent stiffness are obtained. These results highlight that mechanical response emerges primarily from structural design rather than chemistry alone, establishing a framework for mechanically adaptive composite surfaces. Building on this foundation, an emerging direction toward smart material functionalities—specifically magnetic field–induced modulation and piezoelectric self-sensing—is outlined as a potential route for adaptive interfaces in soft robotics.
Özgeçmiş: Dr. Ayşe Çağıl Kandemir received her B.Sc. and M.Sc. degrees in Metallurgical and Materials Engineering from Middle East Technical University (METU). She completed her Ph.D. in Materials Science at ETH Zurich between 2011 and 2016. Since 2018, she has been a faculty member in the Department of Mechanical Engineering at TED University. Her research interests span composite and nanocomposite materials, surface science, thin films, atomic force microscopy, and lithography, with a growing focus on smart materials and functional adaptive surfaces.
Başlık: Improving MRI with RF Engineering
Özet: Magnetic resonance imaging is a medical imaging technique that provides anatomical images with excellent soft-tissue contrast. MR scanners use strong static magnetic fields, magnetic field gradients and radio frequency (RF) fields to create images. Today, the demand for ultra-high-field MR systems is constantly increasing due to their SNR and contrast advantages. New engineering solutions are needed to improve patient safety and imaging performance at these field strengths. This talk will consist of three parts. The first part will summarize our efforts to demonstrate the safety of MRI at 10.5 T, which made it possible to obtain the world's first human brain images at this field strength. The second part will focus on the development of new RF antenna arrays and RF safety validation methods for UHF MRI applications. The last part will present new strategies for safe imaging of patients with metallic implants.
Özgeçmiş: Dr. Eryaman received his PhD in Electrical Engineering from Bilkent University in 2011. After graduation, he worked as a postdoctoral researcher at the MIT-Research Lab of Electronics and the MGH - Athinoula Martinos Center for Biomedical Imaging. He then joined the Center for Magnetic Resonance Research, Department of Radiology at the University of Minnesota, where he is currently a tenured associate professor. He is an awardee of National Institute of Health (NIH)s Pathway to Independence Award-K99 (2016). His current research interests include developing solutions to improve Magnetic Resonance Imaging.
Başlık: Neuromorphic computing - what, why and how
Özet: Since its first use more than 35 years ago, the term "neuromorphic" in the context of computation has carried different meanings to scientists and engineers. The continuous advancement in the field appears to be accelerating in both pace and scope over time, fueled by real-time and low-energy AI thrust. This impetus compels cross-disciplinary R&D collaborations across neuroscience and engineering disciplines more than ever. After reviewing the historical context, we will explore how the growth in neuromorphic computing has significant and exciting implications. We will subsequently examine neuromorphic computing approaches, with emphasis on electronic implementations, supported by examples from our research in resource-constrained systems.
Özgeçmiş: Prof. Dr. Ali Muhtaroglu obtained his earlier degrees in New York, USA (BS'94 in U. of Rochester, MS'96 in Cornell U.). He worked at Intel Corporation for over a decade before joining academics, focusing on low-power mixed-signal circuits and low-energy technologies. He currently leads the Bio-electronics research at the Advanced Health Intelligence and Brain-Inspired Technologies (ADEPT) group at Oslo Metropolitan University. His current research interest includes energy-aware edge machine learning through neuromorphic architectures and circuits.
Başlık: Current Research Activities in the Department of Communications and Information Technologies, ASELSAN
Özet: In this talk, ongoing research activities conducted within the Department of Communications and Information Technologies are presented. First, a bistatic OFDM-based integrated sensing and communication (ISAC) system is examined under a single-target scenario, considering both line-of-sight (LOS) availability and blockage cases. A sliding-window sensing receiver architecture is proposed, through which the inter-symbol-interference (ISI)-free sensing range is extended beyond the cyclic prefix (CP) duration by exploiting pilot symbols embedded in the time–frequency grid. The performance of the proposed architecture is evaluated in terms of range and velocity estimation accuracy and is benchmarked against the Cramér–Rao bounds (CRBs) for the bistatic ISAC setting.
In the second part, a jamming-resilient PUCCH receiver is proposed, in which decoding reliability is assessed through a threshold-based mechanism designed to bound the probability of decoding errors. Furthermore, a channel-estimation-error-aware equalizer tailored for DFT-s-OFDM waveforms is introduced to yield more accurate log-likelihood ratio (LLR) computations. Simulation results demonstrate that decoding reliability can be effectively evaluated without reliance on CRC, even under jamming conditions, thereby improving the robustness of 5G control signaling in contested environments.
Özgeçmiş: Cuneyd Ozturk received the B.S and Ph.D. degrees from the Department of Electrical and Electronics Engineering, Bilkent University, Ankara, Turkey, and the M.S. degree from the Department of Electrical and Computer Engineering, UCLA, Los Angeles, USA, in 2016, 2022, and 2017, respectively. He was a postdoctoral scholar with the Department of Electrical and Computer Engineering at Northwestern University, Evanston, IL, USA, between 2022 and 2024. He is currently a lead engineer at Aselsan Inc., Ankara, Turkey. His current research interests include satellite communication systems, integrated sensing and communication systems, and RIS-aided localization.
Başlık: Safe & Capable: Punyo & Atlas on a Manipulation Spectrum
Özet: In this talk, I'll contrast two approaches to whole-body manipulation on two platforms: Punyo (a compliant upper-body humanoid for home) and Atlas (a powerful humanoid for factories). I'll motivate why whole-body manipulation matters for both humans and robots, then cover two approaches to learning such behaviors: guided reinforcement learning (RL) and behavior cloning. I'll share our findings on the roles of compliance, tactile feedback, domain randomization, and planning in enabling robust whole-body skills on Punyo via guided RL. I'll also give an overview of large behavior models and how they can be adapted to leverage the humanoid form factor for generalist robots. Finally, I'll try to connect these two concepts and open the floor for a collective discussion.
Özgeçmiş: Aykut is a research scientist in the Large Behavior Models (LBMs) Division at Toyota Research Institute, where he is currently developing LBMs for Atlas in collaboration with Boston Dynamics. Previously, he was part of the Punyo team working on contact-rich whole-body manipulation. He received his PhD on planning through contact from Northeastern University in December 2020. Following graduation, he joined Root AI where he worked on mobile manipulation for modern greenhouse harvesting.
Başlık: Model Predictive Torque Control and Fault Diagnosis
Özet: Model predictive control (MPC) is a popular research topic in modern power electronics and has been applied to a large variety of power converters and electric drives. The interest in MPC stems from the advantages it offers: an intuitive and flexible algorithm, fast dynamic response, ease of handling converter nonlinearities, and system constraints.
In this presentation, MPC as applied to the torque control of an induction motor will be presented. Additionally, a novel algorithm that runs as an add-on to the predictive controller to detect and locate inter-turn short circuit (ITSC) faults for the induction motor will be described. ITSC faults are severe and require quick identification to prevent the complete malfunctioning of the machine.
Özgeçmiş: Dr. İlker Şahin earned his B.Sc. (2010), M.Sc. (2014), and Ph.D. (2021) in Electrical and Electronics Engineering from Middle East Technical University (METU), Ankara, Turkiye. He served as a Teaching Assistant at METU from 2011 to 2020 before joining Aselsan in 2020, where he currently holds the position of Team Lead. He also undertook a postdoctoral research role at the University of Edinburgh’s Institute for Energy Systems in the U.K. from 2022 to 2023. His research focuses on high-performance motor drives, predictive control, and fault diagnosis.
Başlık: Toward Real-World Partial Assistance in Lower-Limb Exoskeletons
Özet: Lower-limb exoskeletons promise to enhance balance and mobility, yet real-world daily activities remain challenging due to unexpected perturbations, step-to-step variability, and rapid transitions between locomotion modes. Conventional control strategies, which rely on rigid trajectories or predefined behaviors, often fail to preserve user autonomy or adapt to diverse movement demands.
This talk presents an approach to partial assistance, an assistance paradigm that supports users only when needed while maintaining natural motion elsewhere. I will introduce three components of this framework: a bio-inspired push-recovery strategy that generates adaptive stepping responses; 3D Path controllers that exploit inter-joint coordination, including hip abduction/adduction, for mediolateral balance support; and a personalized, machine learning trained threshold model for real-time locomotion transition detection across different exoskeletons.
Together, these methods demonstrate how integrating dynamic balance mechanisms, adaptive assistance, and user-specific personalization can bring exoskeletons closer to real-world usability, enabling safer, more intuitive, and more natural human–robot interaction in daily living environments.
Özgeçmiş: Zeynep Özge Orhan received her Ph.D. in Robotics, Control, and Intelligent Systems from EPFL, where she conducted her research in the BioRobotics Laboratory, REHAssist Group under the supervision of Prof. Auke Ijspeert and Dr. Mohamed Bouri. She holds a B.S. in Mechanical Engineering from Middle East Technical University (2017) and an M.Sc. in Mechatronics Engineering from Sabancı University (2020). Her research focuses on physical human-robot interaction and control for lower-limb exoskeletons, with an emphasis on partial assistance, balance support, and personalized locomotion strategies. Her work on inter-joint coordination for balance assistance was recognized with the Best Student Paper Award at the IEEE RO-MAN Conference 2025.