Toyota unveils AI Vision Engine for Woven City

Toyota and Woven by Toyota have unveiled AI technologies for the Woven City test environment, led by the Woven City AI Vision Engine foundation model and a broader Integrated Anzen System built around it. The tools are designed to coordinate vehicles, infrastructure and pedestrians through shared sensor data, with the pair planning to deploy the technology beyond Woven City.

The AI Vision Engine draws on camera feeds, mobility systems and user inputs to spot patterns and risks, and ranks among the leading Vision Language Models on the MVBench benchmark. A proof-of-concept is running with UCC Japan, one of the city’s Inventors.

The Integrated Anzen System pairs the vision model with Woven City Behavior AI, which interprets human movement patterns, and Woven City Drive Sync Assist, a driver-assistance layer that reads vehicle and traffic-signal cameras. The three are intended to let vehicles, pedestrians and infrastructure operate as a single coordinated system.

The Woven City Inventor Garage has opened as a prototyping and testing hub at Toyota Motor East Japan’s former Higashi-Fuji stamping plant. AI Robot Association (AIRoA), Daiichikosho, Joby Aviation and Toyota Financial Services have joined as Inventors, taking the tenant list to 24.

Akio Toyoda AI, a model trained on the Toyota Chairman’s leadership approach, was also developed at the site. The Toyota Woven City Challenge accelerator concludes on 23 April, with the winner set to join as an Inventor.

Source: Toyota