A study was performed for the Federal Highway Administration to increase the efficiency of vehicle speed models for the highway performance monitoring system analytical process. Probabilistic and deterministic models developed by the world bank were adapted for conditions in the united states. These models estimate vehicle average travel speed as a function of relevant road and traffic characteristics. This is done by evaluating a set of constraining speed models that consider the influence of vertical grades, horizontal curves, roughness, traffic congestion, and highway type. These models were adapted to conditions in the united states using engineering judgement and limited available data. Although further research is needed to refine these models, the models produce reasonable results and are recommended for use in planning models as a basis for computation of road user costs. More research is needed in this general area from the engineering community. Input from experts in vehicle mechanics, dynamics, and human factors would be particularly helpful in determining driver reactions and behaviour and further developing speed prediction models as a function of road characteristics and vehicle class.
Abstract