Search

​

Left turns are a complicated maneuver at signalized intersections that to accurately model using microscopic (agent-based) simulation requires calibration process that are typically focused on making changes to simulation parameters until field conditions and simulation conditions are comparable. Simulation parameters are often based on hard cut-off points and the definition of zones that, when adjusted, result in a desired driver behavior. This means that, unlike traditional speed input parameters options, defining a distribution for parameters like the critical gap is not an option in common microscopic simulation tools even when gap acceptance behavior is known to be impacted by waiting time, position in the queue, and arguably is likely to be impacted by factors beyond what controlled experiments reveal. Therefore, it is arguably beneficial to incorporate a variation in left turn gap acceptance behavior in agent-based simulations to accurately represent field conditions. This project will contribute to the collective understanding of left turn behavior by monitoring the gap acceptance behavior of individual drivers over time and evaluating the impact that such knowledge can have on the accuracy of agent-based simulations, especially when a probabilistic model that governs gap acceptance are introduced.