基于PSO-BP神经网络的光纤传感器光强补偿

Intensity compensation of optical fiber sensor based on PSO-BP neural network

基于反射式强度调制光纤传感器在测量实验过程中易受周围环境影响造成光源波动以及对探头的欺骗,提出了一种补偿措施,使用粒子群(PSO)优化反向传播(BP)神经网络算法补偿传感器获得光功率值,该算法不仅利用了PSO的寻找粒子群体的最佳位置的搜索性能,还利用了BP算法比较强的局部最优权值阈值搜索性能,粒子群算法优化反向传播神经网络的权值和阈值,从而达到防止反向传播神经网络陷入局部最优的情况。实验中利用光纤探头内圈光纤和外圈光纤接收的光功率值分别对PSO-BP神经网络和反向传播神经网络进行训练,结果表明PSO-BP神经网络的均值误差小于BP神经网络的均值误差,说明其光强补偿的精度更高,该算法能更加有效地减少周围环境影响以及光源波动对光纤传感器光纤接收的影响,有较好实际应用价值。

Because the reflective intensity modulated fiber optic sensor is susceptible to fluctuations in the light source and deception on the probe due to the influence of the surrounding environment during the measurement experiment,this paper proposes a compensation methods that uses particle swarm optimization(PSO)to optimize the back propagation(BP)neural network algorithm to process the optical power value obtained by the sensor.This algorithm not only make use of the search performance of PSO to find the best position of the particle population,but also uses the relatively strong local optimal weight threshold search performance of the BP algorithm.The particle swarm algorithm optimizes the weights and thresholds of the back-propagation neural network,so as to prevent the back-propagation neural network from falling into the local optimum.The PSO-BP neural network and back propagation neural network are trained using the optical power values received by the inner and outer fiber of the fiber probe.Experiments results show that the average error of the optical power value after the improved back propagation neural network processed by the particle swarm algorithm is smaller than that of back propagation neural network,indicating that the proposed algorithm has higher light intensity compensation accuracy,which can more effectively reduce the influence of the surrounding environment and the influence of light source fluctuations on the fiber receiving of the fiber sensor,and has better practical application value.