基于RBF-DRNN应答器传输系统高速适应性评估方法研究  

Research on High-speed Adaptability Assessment of Balise Transmission System Based on RBF-DRNN

作　　者：许庆阳 孟景辉 段贺辉 罗依梦 XU Qingyang;MENG Jinghui;DUAN Hehui;LUO Yimeng(Infrastructure Inspection Research Institute,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China;Beijing IMAP Technology Co.,Ltd.,Beijing 100081,China)

机构地区：[1]中国铁道科学研究院集团有限公司基础设施检测研究所,北京100081 [2]北京铁科英迈技术有限公司,北京100081

出　　处：《铁道学报》2023年第10期71-77,共7页Journal of the China Railway Society

基　　金：中国铁道科学研究院集团有限公司重大基金项目(2021YJ022)。

摘　　要：应答器传输模块(BTM)在不同速度下接收有效报文帧数的准确预测是评估其速度适应性的关键,因此提出一种基于RBF-DRNN的有效接收报文帧数预测方法,用于定量评估传输系统在350 km/h及以上运行条件下的适应性。首先,采用径向基函数(RBF)神经网络建立列车速度与车载设备接收最大、平均、最小比特数之间的非线性回归模型;然后,利用深度递归神经网络(DRNN)建立车载设备接收比特数、误码率、有效率与接收有效报文帧数之间的评估模型;最后,基于RBF模型预测的高速下接收比特数,结合实际误码率、有效率,预测传输系统在更高时速下接收有效报文帧数的变化范围。利用某线路联调联试数据,对模型性能进行测试。结果表明,当列车运行速度达350 km/h及以上时,评估平均误差为0.45帧,最大绝对误差为0.81帧,可有效预测更高速条件下BTM有效接收报文帧数,为应答器传输系统的高速适应性评估提供指导意义。The accurate prediction of the number of effective telegram frames received by the balise transmission system at different speeds is the key to assessing its speed adaptability.Therefore,a prediction method was proposed for the effectively received message frames of BTM based on RBF-DRNN to quantitatively evaluate the adaptability of the transmission system under the operating conditions of 350 km/h and above.First,the radial basis function(RBF)neural network was used to establish the nonlinear regression model between the train speed and the maximum,average and minimum bits received by on-board equipment.Then the deep recursive neural network(DRNN)was used to establish an evaluation model between the number of received bits,bit error rate,efficiency and the number of effective telegram frames received by on-board equipment.Finally,based on the number of received bits predicted by RBF model at high speed,combined with the actual bit error rate and efficiency,the variation range of the number of effective telegram frames received by the transmission system at higher speed was predicted.The performance of the model was tested by using the joint commissioning and test data of a railway line.The result shows that the evaluation average error is 0.45 frames at the train speed of 350 km/h or above,with the maximum absolute error of 0.81 frames,which can effectively predict the number of telegram frames effectively received by BTM under higher speed conditions,and provide guiding significance for the high-speed adaptability evaluation of the balise transmission system.

关 键 词：应答器 高速适应性 RBF DRNN 动态检测 

分 类 号：U284[交通运输工程—交通信息工程及控制]