Machine Learning Assisted FBG-Based Sensors
Many advantages of fiber Bragg grating (FBG) -based sensors over electrical sensors, such as their immunity to electromagnetic interference, high sensitivity, ease of multiplexing electrical and optical sensors, no requirement of additional wires to connect a sensor to the operating unit and many more, have stimulated the idea and implementation in sensing field. FBG-based sensors have been used to measure various physical parameters, such as strain, temperature, pressure, refractive index, vibration, magnetic fields, electric field, moisture, and acceleration among other parameters. The most common difficulty in using FBG-based sensors is differentiating between the changes caused by the strain and temperature, because both fields can change the Bragg wavelength. I propose a machine learning assisted technique to discriminate the changes caused by the strain and temperature. I applied two different types of Machine learning algorithms to predict the strain and the temperature value simultaneously on single measurement of Bragg wavelength. I also applied the proposed technique to an FBG-based current sensor used for monitoring power transmission lines to predict the current and the temperature simultaneously. This proposed approach eliminates the difficulties of the existing discrimination techniques for FBG-based sensors and minimize the cost significantly of the sensor networks.