Self-healing FBG sensor network fault-detection based on a multi-class SVM algorithm
Jinhua Hu, Boying Wang, Kangjian Di, Jun Zou, Danping Ren, Jijun Zhao
Abstract
We propose a three-layer ring architecture with enhanced reconfigurable capabilities for fiber Bragg grating (FBG) sensor networks. The proposed network is capable of self-healing when three fiber links fail. In addition to self-healing, soft faults in the FBG sensors can be detected using a multi-classification support vector machine (multi-class SVM) algorithm. The detection accuracy reached 99%. Additionally, we used an artificial neural network (ANN) reliability estimation model to estimate the reliability of the FBG self-healing network. The results show that the ANN reliability analysis model can accurately estimate the reliability of the architecture at a reasonable cost.