基于深度信念网络的机场噪声数据恢复方法  

Airport Noise Data Recovery Method Based on Deep Belief Network

作　　者：童剑 杜婧涵 胡明华[2] TONG Jian;DU Jing-han;HU Ming-hua(Department of Transport of Jiangsu Province,Nanjing 210001,China;College of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)

机构地区：[1]江苏省交通运输厅,江苏南京210001 [2]南京航空航天大学民航学院,江苏南京210016

出　　处：《航空计算技术》2021年第1期1-4,共4页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(71731001)。

摘　　要：为了缓解机场噪声监测系统中监测点故障期间出现的数据缺失问题,研究了一种基于深度信念网络的机场噪声数据恢复方法,主要应用于机场噪声监测系统中监测点故障期间出现的数据缺失问题。根据故障节点自身的历史时序数据和其高相似度的监测节点的实时数据提取缺失数据的时空相关性。将融合后的数据输入到深度信念网络中提取隐含层特征。通过单层的神经网络预测得到缺失数据的补全值。在机场的噪声数据集中进行的实验表明,所提出的方法相较于传统的数据修复方法具有更高的精度。此外,筛选高相似度的监测节点可以进一步提高补全效果。An airport noise data recovery method based on deep belief networks and temporal spatial correlation is studied,which is mainly applied to the problem of data missing during the failure of nodes in the airport noise monitoring system.First,spatio temporal correlation of missing data based on the historical time series data of the faulty node itself and the real time data of its highly similar monitoring nodes is extracted.Then,the fused data is input into the deep belief network to extract hidden layer features.Finally,the completion value of missing data is obtained through single layer neural network prediction.Experiments conducted on the airport noise data set show that the proposed method has higher accuracy than traditional data recovery methods.In addition,selecting high similarity monitoring nodes can further improve the performance.

关 键 词：机场噪声 机场噪声监测系统 深度信念网络 时空相关性 数据恢复 

分 类 号：V355[航空宇航科学与技术—人机与环境工程]