Modeling Trajectories and Trajectory Variation of Turning Vehicles at Signalized Intersections

Abstract

Information on the trajectories of turning vehicles at signalized intersections can be used in numerous applications, such as movement planning of autonomous vehicles, realistic representation of surrounding vehicle movements in driving simulator and virtual reality applications, and in microscopic simulation tools. However, no proper framework is currently available to realistically model and estimate trajectories of turning vehicles reflecting the intersection geometries, which is critical for the reliability of simulation models. This study explores the applicability of the minimum-jerk principle, which has been initially applied in neuroscience and robotics domains, to model and simulate free-flow trajectories of turning vehicles. The modeling method is validated by comparing model outputs with empirical trajectories collected at several signalized intersections in Nagoya, Japan. The capability of the model in realistically capturing the variations in turning trajectories based on intersection geometry (e.g., intersection angle and turning radius) is also explained. Further, the applicability of the modeling framework at intersections with different geometric features under different speeds and accelerations are also discussed.