Tier IV starts Level 4 autonomy tests across three continents

Tier IV has released software stacks for Level 4 autonomous driving powered by data-centric artificial intelligence (AI), published via Autoware, the open-source autonomous driving platform. The hardware-agnostic stacks support varied system-on-chip (SoC) and sensor configurations and are designed to expand operational design domains through continuous AI model refinement using real-world data.

The release builds on an end-to-end (E2E) architecture published in July 2025 and offers two software configurations: a hybrid system combining perception AI and planning AI using diffusion models to capture temporal environmental changes, and a fully integrated E2E system that unifies perception, planning and control into a single learning pipeline using world model concepts.

To validate the technology, Tier IV has commenced Level 4 driving tests across three regions in collaboration with academic partners. In Tokyo, a Toyota JPN Taxi is used with the University of Tokyo to evaluate urban hub-to-hub journeys. In Pittsburgh, a Hyundai Ioniq 5 is deployed with Carnegie Mellon University on routes including Pittsburgh International Airport. In Munich, a Volkswagen T7 Multivan is used with the Technical University of Munich for urban scenario safety evaluation. Each test run lasts approximately 60 minutes with a safety driver on board.

Demonstration of Level 4 autonomous driving features in Tokyo

Demonstration of Level 4 autonomous driving features in Pittsburgh

Demonstration of Level 4 autonomous driving features in Munich

In a statement, Shinpei Kato, Founder and Chief Executive of Tier IV, said: “Our new data-centric AI models and collaborative MLOps platform provide a common language and a shared foundation for the entire industry.”

Tier IV’s machine learning operations (MLOps) platform handles data-quality validation, anonymisation, annotation and synthetic data generation to support ongoing model development.

Source: Tier IV