TIER IV Launches AI-Based Level 4 Autonomous Driving Software Across Global Test Sites

TIER IV has introduced new AI-based software stacks for Level 4 autonomous driving, expanding testing across Japan, the U.S., and Europe, announced in a press release. The software, built on the open-source Autoware platform, is designed to be hardware-agnostic and supports multiple system-on-chip and sensor configurations.

The release includes two systems: a hybrid model combining perception and planning AI using diffusion models, and an end-to-end (E2E) system integrating perception, planning, and control into a single learning process. Both are available on GitHub through Autoware repositories and are supported by TIER IV’s machine learning operations (MLOps) platform for continuous model improvement.

TIER IV is conducting validation tests with academic partners in Tokyo, Pittsburgh, and Munich. Each test involves Level 4 autonomous driving under local conditions, using different vehicles such as the Toyota JPN TAXI, Hyundai IONIQ 5, and Volkswagen T7 Multivan. Safety drivers will be present per local regulations, though manual intervention is not expected under normal operation.

The company’s Level 4+ initiative aims to expand fully autonomous driving into more complex environments by leveraging real-world driving data and shared development through the Autoware Foundation. This framework enables collaboration among automakers, researchers, and developers to standardize AI-based autonomous driving systems globally.