TIER IV, a pioneer in open-source software for autonomous driving, has been selected by the Japanese government’s New Energy and Industrial Technology Development Organization (NEDO) for an initiative aimed at strengthening industrial competitiveness through the development of a comprehensive AI data platform and ecosystem. Under this project, TIER IV will collect large-scale real-world driving data and generate synthetic data using multimodal AI to cover scenarios that are difficult to capture in the real world. By combining these efforts, TIER IV seeks to create an advanced dataset for autonomous driving AI.
End-to-end AI is gaining traction in the autonomous driving field, and its development relies on vast and diverse driving data. However, collecting real-world data for rare scenarios is both time-consuming and costly.
The project will center on three main areas:
- Large-scale driving data collection: Collection of diverse driving data using multiple vehicle models, and technology development to improve data construction efficiency, such as automated labeling.
- Synthetic data generation using AI: Development of a multimodal generative AI model compatible with multiple sensors to create high-quality training data that covers rare scenarios – such as moments immediately before an accident – and diverse weather and time conditions.
- AI data infrastructure and ecosystem development: Creation of a platform that enables autonomous driving developers in Japan to access high-quality data.
Through this project, TIER IV aims to advance its dataset capabilities while promoting the creation of an AI data platform that reduces the development burden for individual companies. This initiative is expected to enhance Japan’s competitiveness in the autonomous driving sector and accelerate the deployment of safe and efficient autonomous driving systems.
Read Similar News:
Karma Automotive announces Karma Gyesera
