Enhancement of the Multiplexing Capacity and Measurement Accuracy of FBG Sensor System Using IWDM Technique and Deep Learning Algorithm

In this article, we are the first to propose deep learning algorithms for intensity wavelength division multiplexing (IWDM)-based self-healing fiber Bragg grating (FBG) sensor network. A deep learning algorithm is proposed to improve the accuracy of measuring the sensing signal of the sensor system. Furthermore, to increase the total number of FBG sensors multiplexed in the sensor network for multipoint measurements, a multiplexing technique called IWDM is proposed. The proposed IWDM-based ring structure FBG sensor network can also have a self-healing purpose to improve the sensor system's reliability and survivability. However, IWDM has unmeasurable gap or crosstalk problems when the number of FBG sensors increases, which causes high sensing signal measurement errors. To solve this problem, a gated recurrent unit (GRU) deep learning algorithm is proposed and experimentally demonstrated. To prove the sensing signal measurement performance of our proposed algorithm, we test the well-trained GRU model using two cases. The first case is when the spectra of FBGs are overlapped as well as the minimum intensity difference between FBGs is 10%, and the second case is when the spectra of FBGs are overlapped as well as the minimum intensity difference between FBGs is 3% which is a very small intensity difference. From the experimental results, the well-trained GRU algorithm achieves high strain sensing signal measurement performance in both cases compared to other algorithms. Therefore, the proposed IWDM based FBG sensor system using deep learning algorithm enhances the multiplexing capacity and survivability of the sensor system, reduces the computational time, and improves strain sensing signal measurement accuracy of FBGs even when FBGs has very small intensity difference and overlap problem.