This paper presents research under the European Union's road telematics project IN-RESPONSE. The project's main objective is to test and assess innovative incident management policies in several European countries. In France, the Paris Ringway is part of the test site network. Peaks generated by incidents, mainly breakdowns and accidents, are a major restraint on the functioning of this motorway; such nonrecurrent phenomena are today the main source of traffic congestion on the infrastructure. At present, data about detected and processed incidents are centralised and recorded by police operators on an `incident log', and processed to assess the main incident characteristics. A new method has now been developed, which aims at a better knowledge of incident characteristics and uses new telematics tools. For example, a set of fixed video cameras has been installed on a precise zone of the Paris Ringway for automatic incident detection by image processing and for traffic monitoring. The methodology is based on an in-depth analysis of incidents observed by video cameras, and on the design of the incident database. This paper describes the systems, gives examples of its statistical analyses, compares the old and new methods, and discusses detection by machine vision and prospects for incident management.
Abstract