System regulation by the combination of the PCA and the fuzzy logic

This paper proposes a monitoring and regulation approach by combining main component analysis and fuzzy logic. This approach is applied to a motorway history of four variables: traffic density, flow rate, average speed and occupancy rate, and 900 observations. The ACP is used to detect defects by the statistics of the quadratic predictive error SPE and the Hotteling method. The calculation of contributions makes it possible to isolate the defects: to identify defective variables. In our example the defective variables are the flow rate and the average speed. The purpose of fuzzy logic is to develop an access control strategy to eliminate and / or reduce major problems that are beyond the proper functioning.