In our department, the Ph.D. course titled “Special Topics in Mechatronics Engineering, Group 1” (course code MKT6108) will be offered in the Fall 2027 semester. The course addresses the transformation of scientific knowledge and engineering reasoning into clear, executable, and testable structures that can be utilized by small frozen models. It presents an integrated roadmap that begins with the fundamental mechanisms of language models and extends to knowledge architectures, agent-based systems, and the automation of mechatronics engineering processes.
Course Code: MKT6108
Turkish Descriptive Title: Agent-Based Mechatronics Engineering and Knowledge Architectures
Course Content: Structuring scientific knowledge and automating the design processes of mechatronic systems using agent-based structures
Instructor: Dr. Mehmet İŞCAN, Assistant Professor
Course Language: English
Day and Time: Wednesday, 4:00 PM–7:00 PM
Course Format: Online
Course Objective
To equip students with the ability to transform scientific knowledge and engineering reasoning into reusable knowledge and software structures. The course will emphasize the clear definition of assumptions, method selection, calculation steps, and verification criteria within these structures.
To address the end-to-end design of a mechatronic system using controlled, traceable, and testable agent-based systems, and to equip students with the skills to automate modeling, state estimation, control design, system integration, and verification processes.
Scope – Main Topics
Fundamentals of language models: The working principles of language models, their basic generation mechanisms, context management, and the possibilities and limitations of their engineering applications will be addressed.
Scientific knowledge and skill architectures: Scientific resources, expert knowledge, and problem-solving steps will be structured. Reusable skill mechanisms will be developed based on assumptions, feasibility conditions, and resource traceability.
Implementation of agent-based systems: Structured model outputs, computational tools, and agent structures with defined tasks will be addressed. Workflow mechanisms that manage the sequence of operations and hook mechanisms that enforce mandatory controls in specific events will be developed alongside independent verification and state management.
End-to-end mechatronic system design: Requirements definition, dynamic modeling, sensor and actuator models, state estimation, control design, system integration, and performance evaluation will be addressed in an integrated manner.
Scientific experimentation and improvement: Human-supervised solutions will be compared with agent-based execution. Component contributions, error analysis, controlled correction, and retesting processes will be examined. Emphasis will be placed on documenting the applicability limits of these experiments and evaluating them under new task conditions.
Applications and Software
Applications will be based on the transformation of theoretical knowledge into mathematical models, computational methods, and verifiable research software. The work will be conducted in Python and Visual Studio Code environments. The course will emphasize the use of scientific sources and the clear articulation of assumptions, method selections, and validity conditions.
Throughout the semester, the end-to-end design of a mechatronic system will be addressed as a cumulative project spanning 12 weeks. Engineering steps will be executed first using Python and then using agent-based structures to be developed, allowing for a comparison between the two approaches. Skill, agent, workflow, and hook mechanisms will be developed incrementally without using pre-built agent frameworks. Results will be evaluated in terms of scientific accuracy, error analysis, traceability, and reproducibility.
Course lectures, discussions, code, in-code comments, and student technical reports will be in English.
Assessment
Assessment will be based on in-semester research and practical work, a phased design project, technical reports, a final-semester project, and individual scientific presentations. Scientific justification, engineering accuracy, software quality, and the adequacy of verification efforts will be the primary criteria. Students will be expected to explain their design decisions and the validity limits of the results they obtain.
Students interested in this course are strongly encouraged to select it during the registration period.