Wouter J. Schakel, Victor L. Knoop and Bart van Arem
A proposed lane change model can be integrated with a car-following model to form a complete microscopic driver model. The model resembles traffic better at a macroscopic level, especially regarding the amount of traffic volume per lane, the traffic speeds in different lanes, and the onset of congestion. In a new approach, lane change incentives are combined for determining a lane change desire. Included incentives are to follow a route, to gain speed, and to keep right. Classification of lane changes is based on behavior that depends on the level of lane change desire. Integration with a car-following model is achieved by influencing car-following behavior for relaxation and synchronization, that is, following vehicles in adjacent lanes. Other improvements of the model are trade-offs between lane change incentives and the use of anticipation speed for the speed gain incentive. Although all these effects are captured, the lane change model has only seven parameters. Loop detector data were used to validate and calibrate the model, and an accurate representation of lane distribution and the onset of congestion was shown.
Full article available at TRB.
Journal: Transportation Research Record: Journal of the
Transportation Research Board
Publisher: Transportation
Research Board of the National Academies
ISSN: 0361-1981
(Print)
Subject: Transportation
Issue: Volume 2316 /
2012 Traffic Flow Theory and Characteristics 2012: Driver Behavior;
Pedestrian and Simulation Modeling, Vol. 2
Pages: 47-57