AI Model Predicts Driver Behavior at Traffic Signals

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Researchers have developed a new AI model that predicts driver behavior at traffic signals by analyzing decision-making and trajectory during light transitions. This system aims to reduce accidents caused by red-light violations and harsh braking.

Key facts

• A new AI model analyzes driver decisions and trajectories during traffic light signal transitions.
• The model aims to mitigate accidents related to red-light violations and abrupt braking.
• A dataset of 449 real-world approach runs was collected for the study.
• Required deceleration was identified as a key predictor of stop-go decisions.
• The model achieved 0.49m/s² acceleration mean absolute error and 0.62m distance mean absolute error.