改进的高速铁路安全状态协同监测数据处理方法研究  

Study on Improved Data-processing Method of High-speed Vehicles' Safety Monitoring Based on Cooperative-sensing

作　　者：王文韬[1] 

机构地区：[1]北京交通大学电气工程学院,北京100044

出　　处：《计算机测量与控制》2014年第12期3853-3856,3873,共5页Computer Measurement &Control

基　　金：铁道部重点项目(2011Z001-D);轨道交通控制与安全国家重点实验室自主课题(RCS2009ZT003);中央高校基本科研业务费专项资金(2011YJS213;2012YJS113)

摘　　要：将无线传感器网络协同感知方法,应用于高速铁路固定设施状态监测,具有诸多优势,而协同感知数据的处理方法尤为关键;数据融合方法是协同感知方法在数据处理层面的理想选择;在分布式融合场景下,改进的动态误差平方求权系数法,可以充分利用无线节点的存储能力,使得数据波动性减小、一致性增强;在集中式融合场景下,将求取权系数的相关矩阵法从标量推广至多维向量,使得多维向量数据的融合方法得到丰富;仿真实验表明,改进方法下数据准确性、稳定性均优于原有算法;同时,分析比较了选取不同距离函数指数对多维向量相关矩阵法的影响。Method based on cooperative-sensing has many advantages on safety monitoring of high speed vehicles,meanwhile,to handle cooperative-sensing data is crucial.Data fusion is a good way on cooperative-sensing data-processing level.Under distributed circumstances,improved dynamic square error way on calculating weight-coefficient was proposed.As a result,data volatility is reduced,and the consistency is enhanced.Under concentrated circumstances,correlation matrix method was extended to vector aspect.Finally,simulation analysis testifies that the performance has promoted by new algorithms,and algorithm capability with different exponent of distance function is compared as well.

关 键 词：多传感器 协同感知 数据融合 权系数 距离函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]