Optimization of Signalized Intersections: Analyzing Autonomous Vehicle Behaviors Through Data-Driven Simulations

Abstract

Autonomous vehicles (AVs) present a transformative opportunity to enhance traffic flow, particularly at urban intersections where delays are most frequent. This study investigates how different AV driving behaviors and penetration rates affect traffic efficiency at signalized intersections. Using a microscopic simulation model in PTV VISSIM, the research centers on a four-way intersection in Balgat, Ankara. Five AV driving behaviors—cautious, normal, aggressive, platooning, and mixed—are modeled under various signal cycle lengths. The simulation’s accuracy was ensured through calibration and validation with real-world traffic data. The findings reveal that the integration of AVs can significantly improve traffic flow, with aggressive and platooning driving behaviors achieving the most notable reduction in vehicle delays, particularly at shorter cycle lengths (60–70 s). Increased AV penetration rates amplify these positive effects, reducing delays and queue lengths in all tested scenarios. In contrast, cautious AV behaviors led to more significant delays, highlighting the importance of intelligent AV driving strategies for optimizing traffic management. The results underscore that optimizing signal cycle lengths with AV integration can reduce congestion and improve urban traffic flow. While the study demonstrates the potential of AVs to enhance urban traffic management, it also stresses the need for real-world validation and the development of adaptive traffic signal systems capable of accommodating diverse driving behaviors. These insights offer urban planners and policymakers valuable guidance on integrating AVs into current infrastructure to create more resilient and efficient transportation networks. © 2025 Elsevier B.V., All rights reserved.