Research

The purpose is to develop an AI-based prediction model for assessing safety threat levels/situations based on static/dynamic big data collectable on vehicle driving segments (link/node), ultimately aiming to develop a vehicle-centered algorithm for assessing actual safety-threatening situations

ㅇ Identification of risky driving situations and hazardous segments (geometry) based on real-world data - Construction of AI model training datasets through the combination of multiple data sources (CAN data, spatial information data, traffic accident data, etc.) - Development of an AI model for deriving risky driving situations through real-world data analysis - Derivation of use cases that trigger/do not trigger risky driving situations based on AI modeling ㅇ Simulation-based validity analysis and verification of traffic accident contributing factors - Simulation implementation and validity analysis of triggering/non-triggering risky driving use cases - Verification of risky driving contributing factors (segment characteristics & vehicle behavior) and prediction models