CAJ | 학술논문

为了得到列车行进过程中准确的轨道不平顺状态，提出了一种基于以太网的轨道平顺度车载实时动态监测系统，分析了系统的结构和检测原理，通过惯性基准法利用激光位移传感器和振动加速度传感器对轨道不平顺度数据进行采集，基于两传感器数据融合时间不同步、数据率不一致的特点，提出了利用正则化RBF （径向基函数）网络拟合的算法，该方法可以将不同传感器测得的不同采样周期的数据对准到同一时间点上，得到轨道不平顺结果。将拟合值和真实值进行比较，正则化RBF拟合和常规方法拟合的平均相对误差分别为1.9％和3.7％，均方误差分别为0.41和2.35，相比之下，正则化RBF网络拟合效果更好。结果表明该系统能够更真实的反应轨道的动态不平顺。
위료득도열차행진과정중준학적궤도불평순상태，제출료일충기우이태망적궤도평순도차재실시동태감측계통，분석료계통적결구화검측원리，통과관성기준법이용격광위이전감기화진동가속도전감기대궤도불평순도수거진행채집，기우량전감기수거융합시간불동보、수거솔불일치적특점，제출료이용정칙화RBF （경향기함수）망락의합적산법，해방법가이장불동전감기측득적불동채양주기적수거대준도동일시간점상，득도궤도불평순결과。장의합치화진실치진행비교，정칙화RBF의합화상규방법의합적평균상대오차분별위1.9％화3.7％，균방오차분별위0.41화2.35，상비지하，정칙화RBF망락의합효과경호。결과표명해계통능구경진실적반응궤도적동태불평순。
In order to monitor accurately the process of the train track irregularity, this paper propo-ses a real-time dynamic monitoring system for monitoring the track smoothness based on Ethernet by analyzing the structure and detection principle of the system.The track irregularity data is collected by using the laser displacement sensor and vibration acceleration sensor.Since the time and sam-pling rate are not synchronized, regularization fitting algorithm of RBF( Radial Basis Function) net-work is proposed.This method can adjust the different sampling cycle of data from the two sensors to the same time and the track irregularity results can be obtained.Comparing the fitted values and ac-tual values, the average relative error of regularized RBF fitting and the conventional methods is 1.9%and 3.7%, mean square error is 0.41 and 2.35 respectively, showing that the system is more realistically feedback to the reaction dynamics of track irregularity.