Using decision tree regression for estimation of birefringence in mode-locked fiber lasers

BAĞCI M.

APPLIED PHYSICS B-LASERS AND OPTICS, cilt.131, sa.7, ss.1-9, 2025 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 131 Sayı: 7
  • Basım Tarihi: 2025
  • Doi Numarası: 10.1007/s00340-025-08481-4
  • Dergi Adı: APPLIED PHYSICS B-LASERS AND OPTICS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, Chemical Abstracts Core, Chimica, Communication Abstracts, Compendex, INSPEC
  • Sayfa Sayıları: ss.1-9
  • Marmara Üniversitesi Adresli: Evet

Özet

The birefringence in a mode-locked fiber laser cavity varies randomly, and it can dramatically affect mode locking performance. Uncontrollable and unmeasurable fiber birefringence variations make it necessary to adjust the nonlinear polarization rotation (NPR) to achieve mode-locking. Therefore, recognizing the cavity birefringence is of critical importance for algorithms designed for adaptive control and self-tuning of the NPR in mode-locked fiber lasers. In this study, a machine learning procedure based on the decision tree regressor (DTR) is proposed to estimate cavity birefringence. It is demonstrated that the proposed birefringence recognition procedure can estimate cavity birefringence with high recognition rates in a reasonable computation time on both well-aligned and mis-aligned datasets. Thus, DTR-based birefringence recognition approach can be effectively utilized by existing and prospective adaptive control and self-tuning algorithms.