Data-Driven Evaluation Methodology for Active Traffic Management Systems Utilizing Sparse Speed Data

Abstract

Many active traffic management systems and transportation systems management and operations strategies have been evaluated for safety based on crash reduction over time. These long-term studies are effective in showing the safety benefits of new systems, but do not often quantify other factors such as travel time, the extent of congestion, or the environmental impacts. Building on previous research into spatiotemporal interpolation of speed data, this methodology developed a mathematical representation of the speed and acceleration potential of the traffic stream given the sparse speed data from point sensors. This high-resolution estimate of traffic state could be used to construct trajectories of vehicles the could include data on vehicle speed and acceleration at each location in space and point in time. This methodology is general, and the trajectories could be used to evaluate traffic flow in several different ways. In the case study provided, trajectories were used to evaluate the ability of a queue warning algorithm to detect and warn drivers about unsafe conditions. Other potential applications include utilizing these trajectories to calculate fuel consumption, travel times, and speed variability to determine how new systems affect fundamental traffic characteristics.