TOKYO, Aug. 26, 2026 /PRNewswire/ -- TIER IV, the pioneering force behind open-source software for autonomous driving, will showcase a solution for software-defined vehicles (SDVs) that integrates AI, data, and computing at Automotive World 2026, to be held at Makuhari Messe from September 9 to 11, 2026. As part of its exhibition, TIER IV will introduce the autolabeling function of its Co-MLOps platform, a data-sharing platform, as well as its reference E2E AI model, an end-to-end (E2E) autonomous driving AI model developed using data processed through the platform. The demonstration will run the reference E2E AI model on an automotive computer, showcasing an integrated SDV solution that covers the entire process, from data collection and processing to the development of the E2E autonomous driving AI model and its deployment on an automotive computing platform.

Toward the development of autonomous driving systems for passenger vehicles

As the use of AI in autonomous driving technology continues to advance rapidly, high-quality and diverse datasets with accurate labeling are becoming increasingly important for enabling AI to correctly interpret its surroundings. Equally important is an efficient framework for developing autonomous driving AI models using these datasets. To deploy autonomous driving AI models in mass-produced vehicles, including passenger vehicles, it is also essential to run these models efficiently on automotive computing platforms.

To accelerate the development of autonomous driving AI, TIER IV launched the Co-MLOps platform in January 2024 and has been working with partners to collect and share driving data. TIER IV is working toward an integrated SDV solution that combines AI, data, and computing by automatically labeling driving data collected through the platform, efficiently developing the E2E autonomous driving AI model using the labeled data, and establishing an environment for deploying the model on an automotive computing platform.

Key features of the autolabeling function

Using driving data as training data for autonomous driving AI models requires the collection of large volumes of driving data and accurate labeling to enable AI to correctly interpret the data. The autolabeling function automatically labels a wide range of objects and background elements in driving data, including vehicles, pedestrians, road surfaces and structures, so that they can be accurately recognized by AI.

The function can generate millions of labels instantly and with consistent quality, enabling efficient data processing even when new regions or sensors are added or the volume of data increases. By automating the enormous amount of labeling work that has traditionally been performed manually, the function is expected to reduce development costs and accelerate the development of autonomous driving AI models.

To improve the accuracy of autonomous driving AI models, TIER IV supplements difficult-to-collect driving data, such as rare scenarios directly associated with collision risks and data from adverse weather conditions, with synthetic data generated using NVIDIA Cosmos. The autolabeling function is also used to label this generated data.

For more details on the autolabeling function and the use of NVIDIA Cosmos to supplement driving data, please see TIER IV's tech blog, Building a dataset foundation for autonomous driving with NVIDIA Cosmos.

Key features of the reference E2E AI model

The reference E2E AI model does not rely on high-definition maps and uses only images captured by automotive cameras to perform a range of functions, from a bird's-eye-view understanding of the surrounding environment to trajectory generation, using a single neural network. The model supports 3D road and object recognition, the generation of occupancy maps that determine whether obstacles are present in each cell when the environment captured by camera images is divided into a grid, and vehicle trajectory prediction. Its architecture is also well suited to automotive computing, making it easier to deploy on actual system-on-chips.

The model is developed through an agentic AI-driven development process. Humans specify what to build, while an AI agent takes the lead in developing the E2E autonomous driving AI model, incorporating automatically labeled driving data, data transformation and cleansing, implementation optimization, and model management. The agent autonomously runs the cycle of training, evaluation, and improvement. By having agentic AI lead the development process, the model is designed to improve development efficiency and enable low-cost, rapid prototyping of autonomous driving AI models.

TIER IV is continuing to develop the reference E2E AI model and plans to make it available to partners participating in the Co-MLOps project.

Demonstration at Automotive World 2026

TIER IV will demonstrate technology that automatically labels various objects and background elements in driving data using the autolabeling function of the Co-MLOps platform. The demonstration will also run the reference E2E AI model, developed using the labeled driving data, on an NVIDIA Jetson Orin, where the model will use only images from automotive cameras to perform the entire process from a bird's-eye-view understanding of the surrounding environment to trajectory generation.

By integrating large-scale driving data collection and processing through the platform with the continuous development and improvement of the E2E autonomous driving AI model and its deployment on the automotive computing platform, TIER IV is advancing the development of an SDV solution designed to support automakers and suppliers in the development of autonomous driving systems for mass production in passenger vehicles.

About TIER IV

TIER IV stands at the forefront of deep tech innovation, pioneering Autoware, open-source software for autonomous driving. With a comprehensive suite of platforms and services built around Autoware, TIER IV provides everything from software development and vehicle procurement to operational support. Through the Autoware ecosystem, TIER IV works with partners worldwide to shape the future of intelligent vehicles with open-source software, aiming to create mobility that is safer, more sustainable, and accessible to all.

Media contact
pr@tier4.jp 

The word marks and logos of TIER IV are trademarks or registered trademarks of TIER IV, Inc. in Japan and other countries. All other company names, product names, service names, and logos referenced herein are trademarks or registered trademarks of their respective owners.

Autoware is a registered trademark of the Autoware Foundation.

View original content to download multimedia:https://www.prnewswire.com/news-releases/tier-iv-to-showcase-integrated-ai-data-and-computing-solution-for-sdvs-at-automotive-world-2026-302859969.html

SOURCE TIER IV, Inc.