August 25, 2025Scientific ReportsOpen Access

Improving intelligent perception and decision optimization of pedestrian crossing scenarios in autonomous driving environments through large visual language models

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This approach improves decision optimization and risk assessment in pedestrian crossing scenarios, indicating safer autonomous vehicle interactions.

Achieved a perception accuracy of 93.05%, demonstrating effective processing of pedestrian crossing scenarios.
The system utilized visual language models to analyze scenario data, predicting collision risks accurately, while offering actionable safety recommendations.
Through standardized prompts and structured data, the model enhances automated decision-making for improved roadway safety.
This innovative framework outperforms traditional methods, integrating visual perception with reasoning for complex traffic environments.

Xiao et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

This approach improves decision optimization and risk assessment in pedestrian crossing scenarios, indicating safer autonomous vehicle interactions.