What is the TUAT Near-Miss Database?
What is the Near-Miss Database?
The Near-Miss Database, collected from dashcam recordings, is a powerful tool for traffic safety education and the development of preventive safety and autonomous driving technologies. In 2010, the management of this database was transferred to Tokyo University of Agriculture and Technology from the Society of Automotive Engineers of Japan.
To sustain this initiative, a membership-based research group, the "Near-Miss Data Research Association," composed of automotive companies and related organizations, was established to support ongoing operations and advancements in traffic safety.
Database in Numbers
Scale of the Database
220,000+
Near-miss data samples
165,000+
Dual-camera data samples
200+
Used taxis
8 cities
Data collection sites
* Statistics as of October 2024
Database Concept
Extraction of Near-Miss Incidents
15-Second Time-Series Data Before and After the Trigger
More Examples from our Near-Miss Database