基于短时LS-SVM的GNSS铁路路基沉降监测数据分析研究

在GNSS(全球导航卫星系统)铁路路基沉降监测中,卫星信号受到周围环境的影响,同时实时监测中数据传输的不可靠性,导致监测结果往往有很强的不稳定性,有效地剔除粗差能提高GNSS监测的可靠性.鉴于最小二乘支持向量机有很强的非线性运算能力与数据拟合能力,建立短时最小二乘支持向量机铁路沉降数据预测模型,在监测中融合该模型,提高监测的有效性.分别运算该模型对山西中南部铁路通道重载综合试验地的3个监测点进行了分析处理,结果表明融合短时最小二乘支持向量机的预测模型有效抑制了GNSS监测中的离散误差,提高了GNSS监测系统的可靠性与稳定性,为铁路路基沉降监测提供了更为有效的手段.

Analysis and Research on GNSS Monitoring Data of Railway Subgrade Settlement Based on Short Time LS-SVM

In the global navigation satellite system(GNSS) railway subgrade settlement monitoring,the satellite signal is affected by the surrounding environment,and the data transmission is not reliable in real-time monitoring,which results in strong instability of monitoring results.Effective eliminating gross errors can improve the reliability of GNSS monitoring.Least squares support vector machine(LS-SVM) has strong nonlinear computing ability and data fitting ability.The prediction model based on the short-time LS-SVM railway subsidence data was established,the effectiveness of monitoring was improved by integrating the model.The model was used to analyze three monitoring points of railway channel load respectively located in southern Shanxi,the results show that the prediction model fused short-time LS-SVM can effectively suppress the discrete errors in the GNSS monitoring and improve the reliability and stability of GNSS monitoring system,which provides a more effective means for railway subgrade settlement monitoring.