A fuzzy expert system for setting Brazilian highway speed limits

Abstract

Speed limits should provide a balance between safety and traffic flow. However, setting a speed limit generally involves selecting a base speed (e.g., operational speed, design speed) and adjusting it with several other variables. For example, in Brazil, the standard guidelines list a series of factors that influence speed limits, but they do not provide a clear methodology to choose a speed limit for a highway segment. Therefore, under current Brazilian practice, the decision-maker is faced with a subject decision that mostly relies on expert judgment. This paper proposes a fuzzy expert system for setting Brazilian highway speed limits. Six input variables are considered in the system. Membership functions and fuzzy rules were extracted from experts’ evaluations of simulated highway scenarios. The experts evaluated the scenarios using linguistic variables and proposed speed limits. Then, a Mamdani fuzzy controller was developed. The outputs of the controller were compared to the expert’s answers for the simulated highway scenarios. The system was further validated considering two case studies of highway sections in Brazil. Results showed that the fuzzy system is able to provide outputs that agree with the experts’ evaluations and with the existing speed limits. The fuzzy controller developed in this study can be used to assist practitioners in setting highway speed limits on Brazilian highways.