A PRACICAL SIMPLE TECHNIQUE TO DETECT ABNORMAL TRAFFIC FLOW IN FREEWAY

The problem addressed in this paper is the design of asimple and applicable approach with consideringmacroscopic model and stochastic discrete variables todetection of freeway abnormal traffic flow like incident,classified congestion, exit of congestion and so on. Thegoal is to design an algorithm that 1) directly use datafrom conventional presence detectors which providebinary information at each point in time, indicating thenumber of passed vehicles and mean time speed; and 2)minimize human operator requirement in detection,classification, and isolation of incident events. Usingautocorrelation of the time series samples of density andflow which are collected from segments with predefinedspecifications is the main technique to detect the trend inflow and density changes if exist. A table of possibilitiesfor flow and density changes in two sequential segmentswill help to detect congestion or any other abnormaltraffic events.