Fastlane: Traffic flow modeling and multi-class dynamic traffic management

Dynamic Traffic Management (DTM) aims to improve traffic conditions. DTM usually consists of two steps: first the current traffic is estimated, then appropriate control actions are determined based on that estimate. In order to estimate and control the traffic, a suitable traffic flow model that reproduces the properties of traffic well must be used. One of the most important properties is that traffic is composed of multiple vehicle classes. While many traffic flow models have been proposed and applied in DTM, most of them do not capture the dynamics of multiple vehicle classes. In this paper, we propose a multi-class traffic flow model, Fastlane, that reproduces the dynamics and interactions of different vehicle classes. It is especially well-suited for short term multi-class traffic control on freeways. We show three applications of Fastlane: traffic state estimation, traffic state prediction and pro-active control.