Quantifying the impacts of traffic incidents on urban freeway speeds

Planning for traffic related congestion during peak periods continues to be one of the most important challenges facing road managers. Congestion may be thought of as either recurrent or non-recurrent. The latter is caused by factors such as incidents, work zones, weather, and special events. Traffic incidents are reported as the cause of 25 per cent of total delays in the US. However, its effect varies from place to place due to the local conditions. Different types of traffic incidents affect drivers' behaviour and the performance of vehicles. In an incident situation, the average headway between vehicles and the speed variability increases. The resultant impact on road speed profile is the main topic of the paper. The paper describes the methodology of extracting the impacts of traffic incidents including duration and delay on traffic speed for an urban freeway network. The analysis of data from a case study in Brisbane is reported here. A section of an urban freeway has been studied in detail using inductive loop detector data and traffic incidents related variables for a period of 12 months. A variety of probability distribution functions were employed in order to test the best model for the duration and delay frequency distribution for each category of incident. The findings of this research will be used to put forward improved predictive delay models and travel time reliability models for urban freeway conditions.