基于BP神经网络的长距离分布式光纤传感高空间分辨率定位算法OACSTPCD
High spatial resolution positioning algorithm based on BP neural network for long-distance distributed optical fiber sensing
为解决基于数字编码移相-积分测距系统中,因发送序列与本地序列之间的真实相位差非1 bit码长的整数倍而影响测量结果准确性这一问题,文中提出一种基于BP神经网络长距离分布式光纤传感高空间分辨率定位算法.该算法根据BP神经网络判断1 bit内细分程度并与传统算法相结合实现动态测距.实验结果表明,基于BP神经网络长距离分布式光纤传感高空间分辨率定位算法的长距离定位精度平均相对误差从0.23降低到0.035,弥补了空间分辨率与准确率受数字码本身局限性的问题,为大规模基础设施的长距离、高精度实时监测提供了可能.
The real phase difference between the transmitted sequence and the local sequence is not an integer multiple of 1 bit code length,which affects the accuracy of the measurement results in the phase-shifting and integral ranging system based on digital coding,so a high spatial resolution positioning algorithm based on BP neural network(backpropagation neural network)for long-distance distributed optical fiber sensing is proposed.In this algorithm,the subdivision degree within 1 bit is judged according to BP neural network and combined with the traditional algorithm,so as to realize dynamic ranging.The experimental results show that the average relative error of the long-distance positioning accuracy of the high spatial resolution positioning algorithm based on BP neural network for the long-distance distributed optical fiber sensing is reduced from 0.23 to 0.035,which makes up the problem that the spatial resolution and accuracy are limited to the digital code itself,and provides the possibility for long-distance high-precision real-time monitoring of large-scale infrastructure.
金琢然
燕山大学 信息科学与工程学院(软件学院), 河北 秦皇岛 066099
电子信息工程
分布式光纤传感移相-积分测距光时域反射BP神经网络空间分辨率定位测距算法
distributed optical fiber sensingphase-shifting and integral rangingoptical time domain reflectionBP neural networkspatial resolutionpositioning and ranging algorithm
《现代电子技术》 2024 (001)
36-43 / 8
评论