A bayesian network approach to traffic flow forecasting

A new approach based on Bayesian networks for traffic flow forecasting is proposed. In this paper, traffic flows among adjacent road links in a transportation network are modeled as a Bayesian network. The joint probability distribution between the cause nodes (data utilized for forecasting) and the effect node (data to be forecasted) in a constructed Bayesian network is described as a Gaussian mixture model (GMM) whose parameters are estimated via the competitive expectation maximization (CEM) algorithm. Finally, traffic flow forecasting is performed under the criterion of minimum mean square error (mmse). The approach departs from many existing traffic flow forecasting models in that it explicitly includes information from adjacent road links to analyze the trends of the current link statistically. Furthermore, it also encompasses the issue of traffic flow forecasting when incomplete data exist. Comprehensive experiments on urban vehicular traffic flow data of Beijing and comparisons with several other methods show that the Bayesian network is a very promising and effective approach for traffic flow modeling and forecasting, both for complete data and incomplete data.

I. Introduction

Urban traffic control systems (UTCSs) and freeway management systems around the world are collecting large amount of traffic condition data every day. Typical data include volume, flow rate, occupancy, and speed. Development of systems that put these data to good use for traffic control and management has become an active area of ongoing transportation research, which is usually referred to as Intelligent Transportation Systems (ITS) [1]. In the research area of ITS, traffic flow forecasting is a very important issue. Reliable analysis of historical trends of traffic flows is an important input to many of the traffic management and control systems in operation and under development. Some well-known systems, such as the Split Cycle Offset Optimization Technique (SCOOT) system and the Sydney Coordinated Adaptive Traffic (SCAT) system, integrated the traffic flow forecasting function as fundamental modules. Without an effective forecasting capability, these systems would not operate smoothly.