Fiber Fence Alarm System Based On Long-distance And Quasi-distributed FBG Sensors

Fiber fence is a kind of distributed or quasi-distributed sensing system using optic-fiber as the sensing medium, and carries on real-time remote security monitoring of the target within it. Fiber fence, based on long-distance quasi-distributed FBG sensor, not only has the features of electromagnetic immunity, anti-eroding and multiplexing, but also takes advantages of FBG technologies such as low cost, precise alarm locating and so on as a novel alarm monitoring system. Therefore, it holds important application prospects and may become one of the mainstreams in intelligence monitoring field.The author has done a lot of pioneering theoretical and experimental work about this challenging and innovative fiber fence alarm system. The progress schedule is: feasibility analysisâ†'related experiments including long-distance sensing experiment..â†'alarm system experimental modelâ†'alarm system solutionâ†'signal processingâ†'software designâ†'experimental verification. The detailed work is listed below:(1) The fiber fence feasibility solution was analyzed and the present situation of fiber fence alarm system was also studied. It shows that the fiber fence based on the long-distance and quasi-distributed FBG sensor system has significant economic and social value.(2) The EDF/Raman hybrid amplification method was proposed to achieve long-distance sensing, with which the FBG signal obtained at 100km had a SNR of 7dB. Furthermore, the performance of long distance sensing system was improved by adding a LPFG into it. The experimental results show that SNR of the near-end FBG signal is improved by 4.4dB, and SNR of four FBG signals at 50km are all higher than 15 dB. It indicates that the FBG fence monitoring system could work under long-distance.(3) Ground and fence monitoring experimental models are proposed; signal features are extracted and alarm solution based on alarm modes is presented.(4) Dynamic subtraction method is proposed to reduce system error and environmental noise. Wavelet soft-threshold de-noising method is utilized to further improve the SNR. We also use the wavelet multi-resolution analysis method to achieve signal segmentation and present a mean for signal shaping.(5) Alarm signal base model and alarm signal base are built. The statistical features and physical features of the alarm signal are extracted, and physical characteristics combination of alarm mode signal is concluded. (6) The method combining the machine learning and fuzzy logic algorithm is used to judge the alarm modes proposed.(7) Fiber fence alarm system software is designed and developed for achieving alarm and partly judgment function in software form.(8) Simulation of experimental tests including alarm and judgment, and field experimental verification, is conducted, proving that such an alarm mode judgment presented in this work can be applied to fence monitoring successfully.