AI-Enhanced Cascaded Fiber-Optic Sensing Architecture for High-Accuracy Multi-Parameter Monitoring

Abstract

This work presents a novel approach for the simultaneous measurement of temperature and strain using a cascaded multi–single–multi (MSM) mode fiber structure integrated with an artificial neural network (ANN). Experimental data were systematically acquired to construct a comprehensive dataset for ANN training and validation. The proposed model demonstrates substantial performance improvement, achieving reductions in root mean squared error (RMSE) by factors of 1290 and 303 for temperature and strain estimation, respectively, compared to the conventional transfer matrix method. In addition, the proposed framework provides enhanced sensitivity. These results highlight the potential of the ANN-assisted MSM configuration for high-accuracy multiparameter sensing applications