Bayesian network for traffic management application: Estimated the travel time

The notion of travel time is a simple and necessary information for commuters, the intelligent traffic management allows to reduce congestion both temporally and spatially. In this context, this article focuses on the problem of estimating the route time where highway traffic is characterized by several macroscopic variables such as flow, intensity and velocity. We propose a probabilistic model based on Dynamic Bayesian Networks (DBN) which helps us to combine these variables. This model represents a technique for analyzing the traffic to obtain a knowledge model that evolves with time. We estimated, predicted and monitoring travel time given a traffic situation. The travel time is an element that is not directly observable so we used the technique of Hidden Markov Model (HMM) to assure the prediction of travel time in different traffic situations, especially in critical cases where the road traffic is limited.