摘要
对高铁运行网络异常信号在线监测提取问题的研究,能够有效提高网络异常检测效率。对高铁运行网络异常信息的提取,需要计算不完整信号点数据的异常概率,获得剩余不完整信号数据的异常可能性,完成异常信号的监测提取。传统方法获得信号所有尺度空间中的小波模极大值序列,将特征向量输入到支持向量机分类器中,但忽略了获得异常数据的可能性,导致提取精度偏低。提出新的高铁运行网络异常信号在线监测提取方法,首先从高铁运行网络中提取出了5种能明显区别高铁运行网络中不同信号类型的有效频域特征作为异常信号在线监测提取的特征指标;其次计算高铁运行网络中不完整信号点数据的异常概率;然后通过分析高铁运行网络中该信号点异常概率能够直接检测出部分异常信号;最后根据计算获得的剩余不完整信号数据的异常可能性,从而实现高铁运行网络异常信号的在线监测提取。仿真测试证明,所提方法提高了高铁运行网络异常信号检测率,降低了漏报率,且异常信号的在线监测提取成功率较高。
This research focuses on a method for on-line monitoring extraction of abnormal signals in high-speed rail operation network. First of all, five kinds of effective frequency domain characteristics from high-speed rail oper- ation network are extracted as the characteristic indexes of abnormal signal on-line monitoring extraction. These char- acteristics clearly distinguish different signal types in high-speed rail operation network. Secondly, the abnormal probability of data of incomplete signal point in high-speed rail operation network is calculated, and then through the analysis of abnormal probability of this signal point in high-speed rail operation network, some abnormal signals are directly detected. Finally, according to the abnormal possibility of remaining incomplete signal data obtained by cal- culation, the on-line monitoring extraction of abnormal signals in high-speed rail operation network is realized. Simulation shows that the proposed method improves the detection rate of abnormal signals in high-speed rail operation network and decreases the false negative rate. The success rate of on-line monitoring extraction of abnormal signals is high.
出处
《计算机仿真》
北大核心
2018年第3期98-101,144,共5页
Computer Simulation
基金
煤矿井下灾后救援网络重构与应急数据流传输关键问题研究(61471361)
关键词
运行网络
异常信号
监测提取
Operation network
Abnormal signal
Monitoring extraction