Research On Recognition Algorithm Of Optical Fiber Intrusion Signals Based On Φ-OTDR

The optical fiber pre-warning system based on Φ-OTDR(Phase-Sensitive Optical Time-Domain Reflectometer)discovers the intrusion event through distributed optical fiber sensors.Different types of intrusion events can be identified by the processing of optical fiber vibration signals.The recognition accuracy is improved while the time consumption is increased extremely.Especially in the application of long-distance monitoring,it is need to identify a large number of optical fiber signals.In consequence,the real-time performance of the system is facing challenges.In this paper,the problem that the accuracy and time consumption of the recognition algorithm cannot be taken into account is studied,and a hierarchical recognition model based on stress reaction for continuous intrusion event is proposed.The recognition algorithm based on neural network is part of the hierarchical recognition model which some problems in training.Hence,this paper go into the stochastic networks and the stochastic configuration network(SCN)which is one of the most state-of-the-art in stochastic neural networks is studied and improved.This paper proposes a hierarchical recognition model based on stress response process for continuous intrusion events.The mechanism of the stress response is introduced,that is,through the staged consumption of different amounts of human energy,the human body to achieve an effective defense against the invasion of viruses.A hierarchical recognition model for continuous intrusion events is proposed,that is,the continuous intrusion events are recognized in stages,and different complexity algorithms are adopted in different stages.The aim is to improve the recognition accuracy as well as controlling the time consumption lower than those complexity algorithms.In order to verify the performance of the model,the concrete complexity algorithm and low-simple algorithm are adopted in the experiment.The experimental results show that the recognition accuracy of the algorithm based on the hierarchical recognition model can reach the level of the complex algorithm,and the time consumption is greatly reduced.In the hierarchical recognition model,neural network is adopted which has some problems in the process of network training.Therefore,this paper applies the SCN to the recognition of optical fiber intrusion signals.The SCN generates hidden layer node parameters with a constraint condition and controls the scale of the network with stop conditions,which avoids the disadvantages caused by iterative optimization and model scale determination.However,in the application of optical fiber intrusion signal recognition,there is a contradiction between the constraint condition and stop conditions.In this paper,the configuration of hidden nodes in SCN is improved to meet the recognition requirements of optical fiber signals.In this paper,a hierarchical recognition model for continuous intrusion events is proposed to improve the accuracy of recognition while keeping time complexity in mind.Aiming at the shortcomings of the neural network in the model,the SCN is studied and improved to make it suitable for identifying optical fiber signals.It is of great significance for recognizing optical fiber signal quickly and conveniently with network model.