Machine Learning Based H2 Norm Minimization for Maglev Vibration Isolation Platform

BOZKURT A. F., Yalçın B. C., ERKAN K.

Intelligent Systems Conference, IntelliSys 2021, Virtual, Online, 2 - 03 Eylül 2021, cilt.294, ss.100-114 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 294
  • Doi Numarası: 10.1007/978-3-030-82193-7_7
  • Basıldığı Şehir: Virtual, Online
  • Sayfa Sayıları: ss.100-114
  • Anahtar Kelimeler: 4-pole hybrid electromagnet, Linear matrix inequalities, Magnetic levitation, Vibration control
  • İstanbul Ticaret Üniversitesi Adresli: Evet

Özet

Recent advancements in computational power technology and wide-usage of high qualified electronic hardware allow many engineers to create their control strategies in data-driven and model-free approaches, such as reinforcement-learning, genetic algorithm and other machine learning topics, rather than classical and modern control approaches. In this study, Q-Learning RL algorithm combined with analytic LMI method has been utilized to solve micro-scale vibration isolation problem as energy efficient as possible.