Johns Hopkins AI Tool SafeTraffic Copilot Aims to Reduce Maryland Crashes
Johns Hopkins University engineers have developed an AI tool, SafeTraffic Copilot, to enhance traffic safety and reduce collisions. The tool, acting as a copilot, adapts to diverse traffic conditions and cultures, offering detailed risk evaluations and reliable predictions.
SafeTraffic Copilot uses large language models to process vast accident data, including road conditions and numerical values. It can predict how changes in traffic infrastructure, such as traffic light timing adjustments, could affect crash frequency. The tool provides confidence scores for its predictions, addressing the black box nature of AI in high-risk settings like traffic safety.
The tool simplifies complex accident data, offering insights to mitigate crashes for infrastructure designers and policymakers. It can adapt to traffic conditions in other countries and cultures, making it versatile for global application.
So far this year, 381 people have been killed in crashes on Maryland highways, with fatalities rising over the past decade. SafeTraffic Copilot will be applied to improve traffic safety in Baltimore City, Baltimore County, and Maryland, though the specific timeframe and locations for its deployment are yet to be detailed.
SafeTraffic Copilot, developed by Johns Hopkins University engineers, uses AI to predict accident changes based on traffic infrastructure adjustments. It provides confidence scores for its predictions and simplifies complex accident data. The tool is set to improve traffic safety in Maryland, addressing the increasing road fatalities in the state.
