查看文章

Traffic congestion in urban areas is a frequently encountered problem in many countries. This problem can be divided into two groups: recurrent and non-recurrent. The first one occurs routinely and is not triggered by a temporary event. However, the latter is generated in response to temporary events, such as bad weather conditions, work zones, traffic incidents or other non-recurring events. Studies have shown that more vehicle hours of delay result from unexpectedly occurring traffic disturbances rather than recurrent network overload during peak hours. Some of the temporary events, such as accidents, breakdowns or any type of obstacles that can cause induced traffic congestions are called the term incident in the traffic. Incident detection can be seen as a significant component of the overall incident management process. In this study, a literature review is presented on various incident detection algorithms and their real-time and simulation based implementations. An integration of flow and crash modelling that might be useful for developing incident detection algorithms is also discussed.