Application of knowledge-based expert systems to incident management on freeways
Freeway incident management consists of a set of activities designed to detect the occurrence of breakdowns, accidents, and other incidents that affect the flow of traffic, and to initiate responses intended to minimize the adverse effects of these incidents. The dynamic nature of incidents and the uncertainty associated with them require solutions based on the expert judgement of individuals familiar with many interrelated subjects, including traffic control, medical, police, and fire emergency response procedures, and hazardous waste containment. Expert systems attempt to store such human expertise and are designed to replicate the decision making capability of an expert. Expert systems can also process information and screen data, forwarding only the most important information to human traffic managers. They also have the capability to incorporate computerized real time data bases, which, for example, can provide traffic flow information, that would not otherwise be readily available.
This dissertation examines the potential of expert systems applications to freeway incident management and proposes a prototype system. This prototype encompasses incident detection, verification, and response, and includes a real-time dynamic network model for motorist diversion. Simulations of hypothetical incidents on the Massachusetts Turnpike illustrate the potential benefits to be derived from the implementation of the system.