摘要: |
System engineering is used to analyze traffic delay on a test track as a dynamic process using microscale time data. Each driver is required to start from rest and stop at a specific location in the shortest time possible. As such, delay is carefully defined and used as a measure of performance. The focus of the study is to identify triggering delay events and explain their origin. Road network design is treated as an explanatory variable because the test track has two cruising zones with different speed limits. In one direction, the drivers are required to accelerate at the trip midpoint, and in the other direction, decelerate at that point. Stochastic models mimic the behavior of a driver of a leading vehicle and a car follower. Study findings, based on the analysis of 400 test track runs, identify two delay event types that warrant action: (1) eliminate speed drops at a nonbottleneck location, and (2) eliminate extremely long vehicle spacing gaps found on rare occasions in cruise zones. Various tools are used to obtain a firm understanding of the who, what, when, and where of driver decision making. They include acceleration, speed, and position trajectories to clarify driver actions at startup when cruising and stopping; time-series averaging to obtain a global view of operations; and histograms to study individual behavior. Stochastic roadway network model assembly is presented. Discussions on how to improve a behavioral car-following network model (bCFNM) computer simulator, data validity testing, and model parameter calibration are presented. |