Estimação de medida locacional fuzzy para contagem volumétrica de veículos em via urbana

Abstract

The volume of urban traffic presents seasonal variations with peak times during the daily period. This scenario has caused congestion and, these variations of the traffic flow influence the average. Thus, this proposal suggests a robust measure locational Fuzzy, which translates to Count Volumetric vehicles (CVV) in pertinences, which are used as weights to the Fuzzy average. In obtaining this locational measure were used FNM and GFNM methods, that using the extension principle, assessed by Euclidean distance and the Mahalanobis Distance, respectively. The calculation of fuzzy robust locational measure will minimize the variability of the flow of vehicles without excluding the observations that were considered outliers. The validation of the suitability of this average strength is measured by the criterion function that must be minimized iteratively, reaching less than that corresponding to the previous iteration. The application was made for hourly intervals of CVV on weekdays to the intersection of Av. Genaro de Carvalho and Street Benvindo de Novaes, place of frequent changes during the time period of the day. The method gave priority to check the process of behavior that could possibly be considered as candidates for possible levels of retention. Fuzzy robust medium considered were those that presented the convergence criteria for the function