Highway incidents, primary contributors to traffic congestion, often cause not only significant delays to the daily roadway users, but also the reliability of transportation systems and even the productivity of the supply-chains of some vital industry sectors. To mitigate the impacts of such incidents and to recover the performance of highway systems as safely and quickly as possible, it is essential that the responsible highway agencies shall operate an efficient system to detect an incident, estimate the required clearance duration, assess the resulting traffic impacts, and then take necessary control actions.

To do so, the most critical task is to have a reliable estimate of a detected incident’s impacts. However, providing the information of time-varying incident impacts to the general public at the desirable level of accuracy is a challenging task due to difficulties in having sufficient data and the complex relations between key factors contributing to the incident impacts. The purpose of this study is to develop a traffic incident management (TIM) support system which is capable of providing robust and reliable information with respect to the estimated clearance duration and capacity drop of a detected incident and its temporal as well as spatial evolution of traffic impact patterns.

The proposed incident management support system consists of two main components, one for estimating the incident duration and the other for computing the resulting capacity drop of the roadway segment plagued by the incident. The first component servers to provide a robust incident duration estimate, using several specially-designed methods to effectively tackle the unique distribution patterns of the incident duration data and their complex correlations among contributing factors, which includes classification model, continuous model, and supplemental rules to first produce an initial interval estimate, and then a point estimate along with the outlier information. The second component is designed to estimate the additional roadway capacity reduction (i.e., capacity drop) due to the lane blockages by incidents and the response operations, allowing the control center operators to assess the spatial and temporal incident impacts on the highway network, and take necessary control actions in a timely manner.

The proposed TIM support system has the following key features: 1) providing the initial estimate of incident duration, based on limited data available at the early stage of incident responses and operations; 2) updating the estimated incident duration with a specially-design process and models when more data become available; 3) implementing an integrated estimation methodology to circumvent the variances due to the unique characteristics associated with recorded incident data (e.g., highly skewed distribution, complex correlations among the explanatory variables, mixed qualitative and quantitative variables, and heteroscedasticity); 4) generating the estimated additional capacity reduction for the highway segment plagued by the lane-closure and lane-changing activities during the incident clearance operations using a reliable and trackable analytical model; and 5) providing a convenient and effective computation process to estimate time-varying incident impacts, such as queues and delays, in the highway network for real-time applications.