FBG Based Optical Weight Measurement System and Its Performance Enhancement Using Machine Learning

Industrial weight measurement is essential to monitoring and optimizing the quality and production cost. Fiber Bragg gratings (FBGs) are among the most widely used fiber optic sensors because they have a high sensitivity to temperature and strain fluctuations, making them ideal for various industrial applications. This paper proposes an optical weight measurement system that includes four FBG based load cells with a detailed analysis of their performance using different machine learning algorithms to improve accuracy and measurement range. Here, the shift in Bragg wavelength caused by strain generated on cantilever type load cell is acquired using an optical interrogator, giving the wavelength encoded FBG data. The data is acquired on PC, where specific Machine Learning models such as Linear Regression (LR), Support vector machines (SVMs), Neural networks (NN), Gaussian Process regression (GPR) are optimally selected and are utilized to directly learn the mapping between changes in load cell (Bragg wavelength) and applied weight. The experimental analysis shows that the GPR algorithm provides more accurate weight measurement than any other standard approach. The model ensures the practicality of the proposed method with improvement in measurement speed and accuracy of the FBG based weight measurement system.