Improved vehicle speed estimation procedures for air quality and planning applications

Estimation of vehicle speeds under various traffic and highway conditions is a high priority for transportation analysts. Speeds are used directly in air quality models to estimate vehicle emissions. Other measures such as vehicle delay and travel rate can be derived from speed estimates and are commonly used as the fundamental measures of urban area congestion. However, most prior research has focused on speed estimation for uncongested (i.e.. unsaturated) traffic conditions. Further, the estimation process is more tenuous for transportation planning applications where very little input data are available.

This research focuses on the development, calibration, and application of a new methodology for estimating vehicle speeds for air quality and other transportation planning applications. The methodology uses data that are readily available to transportation planners and embodies several theoretical and practical concepts missing from traditional speed estimation methodologies. The main features of the approach include: the use of temporal distributions as a basis for developing hourly traffic estimates; accounting for daily variation in traffic by allowing hourly traffic estimates to vary stochastically; use of a capacity drop on freeways after flow has broken down; separate functions to estimate speeds in queuing and free flow conditions; and the use of highway capacity concepts to determine when traffic operates under free flow and queuing conditions as well as a basis for estimating free flow speeds and the extent of queuing. The new methodology has been developed in two forms: (1) a set of simple equations that can be applied to individual highway segments or network links and (2) a detailed method for tracking the speed characteristics of highway segments over time and space.

Application of the methodology at both the network and corridor levels leads to several conclusions. The main conclusion is that as the extent of congestion grows (as defined by traffic in queues), the methodology predicts lower speeds than traditional methods; the more extensive queuing is, the larger this difference. Other conclusions include; traditional speed estimation methodologies produce a wide range of results and do not explicitly consider the effects of queuing; congestion is more appropriately related to AADT/C rather than to V/C; on signalized arterials, signal density and progression were found to be strong determinants of speeds.