Autonomous Vehicle Injuries Could Decrease With Better Street Lighting and Intersection Designs: Study

A new study shows that the type of vehicle, its manufacturer and traffic signal systems are crucial factors that contribute to accidents involving autonomous vehicles (AVs).

Autonomous vehicles, or self-driving cars, use sensors and AI to operate without much, if any, human intervention. However, as AVs have become more popular among consumers, these relatively new technologies still face many unknowns in real-world scenarios.

These concerns have been highlighted by a series of high-profile autonomous vehicle crashes in recent years, which have led to investigations by the National Highway Traffic Safety Administration (NHTSA). These incidents have raised questions about how well AV technology integrates with existing roadway infrastructures and the effectiveness of current safety protocols.

In a study published in Science Direct earlier this month, researchers used advanced machine learning to develop a model that can predict severe crashes involving autonomous vehicles. According to the authors, this new model could help improve road safety by identifying potential accidents before they happen.

In the study, a team of researchers from the University of Texas used a comprehensive dataset from the California Department of Motor Vehicles, which included crash data on 358 AV accidents from 2014 to 2024.

The researchers, led by Sai Sneha Channamallu, designed a model that identifies key predictors of crash severity, which include the extent of AV damage, vehicle type, manufacturer and the presence of traffic signals.

The study found that sub-compact and compact cars were associated with more severe injuries for drivers due to their smaller size and mass. In addition, 72% of AV-related accidents occurred primarily on streets, and 64% took place at intersections. Traffic signals were present at 53% of these intersections.

Street lighting was identified as the eighth most significant predictor of crashes, as poor lighting conditions contribute to more severe accidents and significantly impact crash outcomes.

Researchers emphasized that implementing an advanced model similar to the one they used for crash injury predictions could greatly benefit vehicle manufacturers, urban planners, policymakers and end-users.

The results suggest that improving intersection design, real-time monitoring, adaptive control systems and lighting conditions is crucial for reducing crash severity and enhancing overall AV safety.

The study’s findings show promise in improving roadway safety by using long-term data collection to guide future design improvements.

In addition, integrating the model the researchers used into onboard diagnostics could enable real-time vehicle assessments. This could allow vehicles to take preventative actions, such as adjusting speed or alerting drivers, which could help prevent accidents and enhance overall safety.

“Key predictors of crash severity were found to include the extent of damage to the AV, the manufacturer, the type of vehicle involved, and the presence and type of traffic signals at the crash site,” Channamallu said. “The nature of the collision, intersection geometry, vehicle actions prior to the crash, and lighting conditions were also found to significantly influence crash outcomes. These findings emphasize the importance of robust safety mechanisms, continuous innovations in AV technology, safety measures tailored for different vehicle types, and effective traffic signal systems.”