基于Logistic集成学习的列车MVB网络异常检测方法研究

多功能车辆总线(Multifunction Vehicle Bus,MVB)已经广泛应用于轨道交通车辆,而恶劣的工作环境易造成MVB网络通信性能退化,严重时危及行车安全.在对MVB网络常见故障进行分析的基础上,从MVB物理层和数据链路层中提取网络状态特征,提出了一种基于异质Logistic集成学习的MVB网络异常检测方...

Anomaly Detection for MVB Network Based on Logistic Ensemble Learning

Multifunction vehicle bus(MVB) has been widely used in rail transit vehicles, but the poor working environment may result in the performance degradation of MVB network, which even endangers the driving safety in serious cases. Based on the analysis of common faults in MVB network, with extracting the network state features from the MVB physical layer and data link layer, an detection method based on heterogeneous Logistic ensemble learning to detect MVB network anomaly and avoid breakdown maintenance to the maximum extent was proposed. A MVB network experiment platform was constructed, and multiple sets of fault injection experiments were conducted. The experimental results showed the validity of the proposed method.