期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：16
页码：3870
出版社：Journal of Theoretical and Applied
摘要：Automation has become important to meet the needs and requirements of industries especially the transportation industries. Thus, in a transport context, the aim is to increase the level of traffic flow and to develop monitoring techniques by monitoring the evolution of the performance indicators of a system; The detection of a defect and its diagnosis are of great interest and the prognosis is currently the subject of several in-depth studies. This paper suggests a detection approach in the flow problem regarding the road traffic the rough principal component analysis (PCA). This control technique is considered very effective in the field of surveillance. The PCA is considered an indispensable tool applied by industries with the objective of avoiding or reducing any anomaly that may intervene in the field of operation of the system. . It is applied to a road section of Lille-France with 829 measurements with four variables: traffic density, flow rate, average speed and occupancy rates following the application of a modeling tool of the second order in the Pre-processing phase of the data. The PCA is used to detect defects by statistical predictive square error SPE and method Hotteling T2. Thus before the isolation, the segmentation has become a necessary step to ensure the visualization of classes (in faults and without defects). The calculations of contributions allow isolating faults and identifying faulty variables. In our example, the defectives variables are both the flow rate and the average speed.
关键词：Modeling of second order; Linear PCA; Fault detection; segmentation; Fault isolation
