节点文献
刻纹光纤传感阵列神经网络处理的可行性研究
Feasibility of Neural Network Processing for Etched Optical Fiber Sensor Array
【摘要】 提出了一种新颖的、采用周期性双面不对称刻纹多模光纤组成的传感阵列,可用于机敏材料与结构的状态探测与损伤估计.探索了采用人工神经网络处理其阵列传感信号的可行性,并给出了BP网与Kohonen网的仿真结果.
【Abstract】 A novel fiberoptic sensor array which is made of two-surface asymmetric pe-riodically etched multimode optical fibers is presented. It may be applied to detection ofstructure state such as stress, strain or damage assessment is smart material and struc-tures. The technology of array sensing signal processing in artificial neural network isdescribed. The simulation results of BP network and Kohonen network are also given.
【关键词】 光纤传感;
机敏材料与结构;
人工神经网络;
【Key words】 fiberoptics sensing smart material and structure artificial neural network;
【Key words】 fiberoptics sensing smart material and structure artificial neural network;
【基金】 国家科技预研基金资助项目
- 【文献出处】 传感技术学报 ,Journal of Transcluction Technology , 编辑部邮箱 ,1996年04期
- 【分类号】TP212.14
- 【被引频次】4
- 【下载频次】32