Tokyo police plan AI system to reconstruct traffic crashes
Testing is planned for June 2027, with full operations scheduled for January 2028.
Tokyo’s Metropolitan Police Department plans to use artificial intelligence to reconstruct traffic accidents, combining three-dimensional mapping of crash locations with recordings from nearby security cameras. The system is intended to reduce the need for lengthy inspections that restrict traffic.
The department has included related costs in its budget request for the next fiscal year. It plans to begin testing in June 2027 before moving to full operations in January 2028.
According to the MPD, the technology will draw on an existing mobile mapping system that uses vehicles equipped with cameras to record their surroundings and produce 3D spatial data. Investigators will be able to retrieve vehicle speeds, distances and relative positions on a screen.
Simulations will offer overhead views and perspectives from drivers and pedestrians, helping investigators prepare diagrams and reports.
Current investigations into fatal crashes and hit-and-run cases involve placing vehicles in their crash positions and taking measurements at the scene. These inspections can restrict traffic for two to three hours and sometimes require more than 20 personnel, including traffic controllers.
