Bayesian Reasoning for OD Volumes Estimation in Absorbing Markov Traffic Process Modeling

Traffic became one of the highest nowadays problems. The number of cars on the road is continually increasing and solutions to improve the traffic conditions are mandatory. We cannot prohibit people to use vehicles, but we can help them by using a traffic management system. The purpose of these systems is to update in real-time the timing of traffic signals, to reduce the traffic congestion. ITS (Intelligent Transportation Systems) is the concept associated with these approaches. The purpose of this paper is to bring a new solution to improve traffic conditions. Starting from a microscopic traffic model, we propose an improvement of OD (Origin-Destination) matrix estimation. This estimation will be capable to lead to better green intervals choosing for traffic signals. Our approach brings a new algorithm for traffic management, at traffic OD volumes estimation level, that combines the probabilistic concept of Bayes inference with absorbing Markov process. These estimations can be further used to update the traffic signals timing and phases.