Analysing freeway traffic incident duration using an Australian data set

This paper investigates incident duration and identifies contributing variables for Australian conditions. The paper presents a new framework for comprehensive traffic-incident data mining and analysis towards an incident delay model and travel-time reliability modelling. Twelve months of data were collected, analysed and the results are presented in this paper. The findings suggest that debris, breakdown and multiple-vehicle crashes are the major sources of incidents on freeways. Furthermore, freeway incident duration varied across the types of incident and time of the day, and whether it was a week day or weekend day. However, there were no significant differences in relation to day, week or month of the year. Significant variables on incident duration were identified using an ANOVA test for each type of incident. In addition, the findings of this study reveal a high variance of incident duration within each incident type. A variety of probability distribution functions were employed to test the best model for the duration frequency distribution for each category of incident. Log-normal distribution was found to be more appropriate for crashes, but log-logistic distribution was more appropriate for hazards and stationaryvehicle incidents