A stochastic modeling approach to dynamic prediction of section-wide inter-lane and intra-lane traffic variables using point detector data

Real-time section-wide lane traffic variables such as density and lane-changing are vital to traffic control and management in urban areas. They can be used as decision variables to determine traffic control and management strategies in real time as well as characterize road traffic congestion for further use in advanced traveler information systems. Therefore, developing techniques which provide real-time information regarding section-wide inter-lane and intra-lane traffic variables is an increasingly important task in the area of advanced transportation management and information systems. This paper presents a stochastic system modeling approach to extracting real-time information of section-wide inter-lane as well as intra-lane traffic (e.g. lane-changing fractions, lane densities, etc.) utilizing lane traffic counts detected from point detectors. The proposed methodology consists of three principle elements: (1) specification of system states, (2) system modeling, and (3) recursive estimation. Preliminary test results indicated that the proposed methodology is promising for estimating real-time section-wide inter-lane as well as intra-lane traffic variables based merely on point detector data. The inter-lane and intra-lane traffic information generated by the proposed method can be further used in developing related technologies such as road traffic congestion detection, automatic incident detection, prediction of driver route choices, variable message signs and in-car navigation devices.