Investigating the impact of safety warning system on road user behaviors using vision-based tracking algorithm

Introduction: This study utilizes a non-intrusive before-and-after approach, employing a vision-based tracking algorithm to examine dynamic traffic data collection and monitor the complex behaviors of road users in relation to safety-related environmental modifications.

Method: The proposed methodological framework is employed to assess the effectiveness of a smart crosswalk system integrated with in-pavement LED signal lights, visual and sound warning systems, and electronic billboards in influencing behavioral adjustments among road users and reducing their potential crash risk. We conducted a thorough analysis of road user trajectories extracted from video recordings captured before and after the installation of the smart crosswalk system.

Results: Our findings provide compelling evidence that the implementation of these innovative interventions significantly enhances pedestrians' situational awareness, thereby improving overall pedestrian safety. However, our models indicate a potential increase in vehicle speed following the physical modifications.

Conclusions and practical applications: Based on these findings, we propose practical guidelines for the appropriate installation of smart crosswalks to achieve the dual objectives of promoting safe crossing and enhancing the perceived convenience for road users.