Accurate Modeling of Distributed Bragg Reflector Laser Power and Wavelength Using Gaussian Process Regression

• Published in: Photonics Vol. 10; no. 2; p. 193
• Main Authors: Yue, Ziqian, Cao, Li, Wang, Dawei, Yuan, Ziqi, Li, Jiajie, Chen, Baodong, Zhai, Yueyang
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.02.2023
• Subjects: Atomic physics; Bragg reflectors; DBR; DBR lasers; Efficiency; Frequency control; Gaussian process; Gaussian process regression; Gaussian processes; Laser outputs; Lasers; Machine learning; Methods; Normal distribution; Physics; power; Radiation; Random variables; Root-mean-square errors; Wavelength
• ISSN: 2304-6732; 2304-6732
• DOI: 10.3390/photonics10020193

Distributed Bragg reflector (DBR) lasers are widely used in many physics experiments. However, regarding the power and frequency control of DBR lasers, obtaining complete and accurate output characteristics is challenging due to the need for general and accurate quantitative models. In this study, we propose and validate a method based on Gaussian process regression to quickly and accurately establish the DBR laser output power and wavelength model. Two models are developed to describe the output power, wavelength, input current, and temperature. The findings show that our power model explains the laser’s power change from the current threshold to the maximum value more precisely, with a root mean square error (RMSE) of 0.16921 mW, less than one-fifth of that of the classic power model. Moreover, our wavelength model is feasible for accurately describing the laser wavelength with a RMSE of 4 × 10−4 nm. This study can improve DBR laser power and frequency control efficiency and precision.