Research On A Robustness Evaluation Model For Optical Fiber Sensor Networks And Its Application

Optical fiber sensor networks (OFSNs) have attracted increasing attention withthe development of optical fiber sensing technology, especially in applications such asmajor security warning and structural health monitoring. However, OFSNs are facingthe problem of a lack of systematic evaluation criteria to assess network performance.Research on robustness for OFSNs thus far is merely focused on designingarchitectures that have high robustness. No universal method for quantitativerobustness evaluation has been proposed. Thus it is difficult to tell how much onetopology is superior to another in terms of robustness. It is crucial to establish a modelto accurately evaluate these OFSNs. This paper explores into the metric of robustnessevaluation for OFSNs. We propose a method to assess robustness and a universalquantitative robustness evaluation model for OFSNs is established. The model definesrobustness as the mathematical expectation of the monitoring coverage ratio, whichhas taken into account the performance under all possible network states and theprobability of each state. With this robustness evaluation model, we are able tocompare the robustness of different OFSNs quantitatively and decide the optimumparameter in the designing of OFSNs for specific applications. On the basis of thismodel, we have described in detail how to calculate the robustness of several basicnetwork topologies, including line, ring, star and bus topologies, respectively, andshowed the robustness expression of these four topologies. Using the Support VectorMachine (SVM) and the Monte-Carlo method, several influencing factors, such as theattenuation coefficient, the threshold, the number of optical fiber sensors, the topologyand the monitoring area, are investigated for their impact on the robustness of thenetwork both through simulations and experiments. Finally, experimental verificationis performed and the robustness evaluation model is applied to several specificapplications to verify the validity of the model. The work in this paper is helpful forthe estimation of adaptability and applicability of OFSNs, and has some degree ofguiding significance in the network design.Main work of this paper: 1. We have proposed for the first time a method to assess robustness andestablished a universal quantitative robustness evaluation model for OFSNs, whichprovides a feasible solution to the fuzzy problem in the robustness evaluation forOFSNs.2. Continuous and distributed measurement of the entire monitoring area isachieved by only using small number of discrete optical fiber sensors and the SupportVector Machine. And this is proved both through simulations and experiments.3. Based on the robustness evaluation model, the robustness expression of thesefour topologies are theoretically derived. Simulations and experiments are performedto see how environmental settings and influencing factors, including attenuationcoefficien α, threshold γ, the topology, the number of sensors, impact the networkrobustness.4. The error analysis of the robustness evaluation model is carried out through theMonte-Carlo method. The robustness of four-sensor network composed of line, ring,star, bus and daul-ring topology are calculated respectively. Result shows that thedaul-ring topology has higher robustness than the other four topologies.