Vulnerable roadway users safety project
The California Department of Transportation (Caltrans) is evaluating if generative artificial intelligence (GenAI) can help better identify locations that need infrastructure improvements to enhance vulnerable roadway user (VRU) safety.
The opportunity
Caltrans’ vision is to end road fatalities and serious injuries on California roadways by 2050. Around 25% of crashes involve VRUs. According to the California Strategic Highway Safety Plan’s 2023 VRU Safety Assessment, one in 4 fatal or serious injury VRU crashes involve pedestrians and bicyclists. This emphasizes the need to focus on VRU safety.
Right now, most safety processes use a reactive approach. Crashes are analyzed after they happen. This involves prioritizing infrastructure improvements at High Crash Concentration Locations (HCCLs) and selecting appropriate treatments to remedy the locations.
Caltrans lacks comprehensive data inputs and technology to analyze large volumes of structured and non-structured datasets. This prevents them from proactively identifying VRU safety locations.
How Caltrans is exploring GenAI
Caltrans wants to test if GenAI can be harnessed to transform traffic crash data into actionable insights to proactively identify locations needing safety enhancements. This will help Caltrans to improve VRU safety and aid the state in making necessary investments with more efficacy, especially in underserved areas.
Caltrans is using GenAI to analyze several datasets to improve the identification of potential locations and appropriate safety treatment recommendations. They include:
- Historical crash information
- Vehicular traffic volume
- Bike and pedestrian volume
- Highway inventory
- Equity and demographic metrics
- Point of interests (like hospitals or schools)
- Connected vehicle data
The goal is to find regional trends and potential HCCLs. Transportation staff intend to use GenAI insights to better address VRU safety issues.
Learn more about Caltrans’ VRU project on Cal eProcure.