Development of Prototype Expert System for Roadside Safety

Abstract

Collision of vehicles with fixed objects adjacent to the roadway has been a concern of traffic engineers for many years. One solution to prevent or reduce the number of accidents of this type is to install a traffic barrier to shield the objects from being hit directly. However, because the barrier itself is a fixed object, the question is whether a barrier is more effective at a particular location than no barrier at all. If a barrier is needed, where should it be located and what type of barrier would satisfy the design criteria? Issues of this nature are ill‐structured and difficult to address using normal computing technology since many factors contribute to the problem. For roadside safety, the primary factors are diverse and interrelated. This paper presents a demonstration prototype knowledge‐based expert system (Roadside) for roadside‐safety analysis. In particular, this system is developed to evaluate whether a traffic barrier is necessary for a particular site in analysis. Development of this demonstration prototype shows an expert‐system approach for roadside safety analysis is feasible. The artificial‐intelligence (AI) programming language Prolog is used for developing the system. The use of Prolog provides greater flexibility and adaptability in developing the system. However, much more programming time is required to develop the expert system.