The problem of recommending a suitable speed limit for roads is important for road authorities in order to increase traffic safety. Nowadays, these speed limits can be given more dynamically, with digital speed regulation signs. The challenge here is input from the environment, in combination with probabilities for certain events. Here we present a decision support model based on a dynamic Bayesian network. The purpose of this model is to predict the appropriate speed on the basis of weather data, traffic density and road maintenance activities. The dynamic Bayesian network principle of using uncertainty for the involved variables gives a possibility to model the real conditions. This model shows that it is possible to develop automated decision support systems for variable speed regulation.
Paper Id: 0022
Presented in Helsinki