Software Design Of Distributed Optical Fiber Sensor Based On Embedded System

Distributed optical fiber sensors,as a new sensing technology,are used to detect the temperature and strain information of the external environment,and are widely used in various fields by virtue of its many advantages.Based on the research that has been made by our laboratory in recent years,a set of integrated distributed optical fiber sensing system is designed and developed in this paper.Based on Brillouin scattering sensing technology,the temperature and strain information of the environment are detected according to the linear relationship between Brillouin frequency shift(BFS)and temperature and strain.The sensing system is divided into two separated sub-systems,which are server and client.Meanwhile,the server based on embedded platform has advantages of flexible,low cost and high stability etc.This paper focuses on the design and development of the server system,which uses ARM+DSP dual-core processor as the embedded platform.The raw data of the sensor system are preprocessed by FPGA at a high speed to generate low-speed data,which are processed by DSP with a specific algorithm to obtain physical quantity information such as temperature and strain of the real environment.And these information is displayed by ARM with an embedded Linux system for human-computer interaction.At the same time,the results are encapsulated in XML format and sent to the client thr-ough the network module.In this paper,the data acquisition and communication process of the sensor system is developed and tested,and the results are in line with expectations.In addition,this paper focuses on the fast algorithm of calculating BFS from original data of distributed sensor system.Based on the shortcomings of high time complexity of the current mainstream curve fitting algorithm,we proposed a novel fast interpolation algorithm based on cross-correlation technique to estimate the BFS.The performances of frequency accuracy and time complexity of the proposal are investigated through simulation and experiment.The results demonstrate that the proposal provides the same frequency accuracy compared to the fitting methods.But the proposal has much less computational complexity,so the method of this study can substitute the standard fitting process,which is of great value for the distributed dynamic measurements and fast processing of the sensor system.