AD/ADAS Country-Based Virtual Validation Using Real-World Data

Accurate 3D simulation models that can represent country-specific features are key to accelerating the development and virtual validation of innovative ADAS/AD functions. This presentation from IAV illustrates how to enhance 3D models with real-world test data by applying AI methods with MATLAB^®, Simulink^®, and RoadRunner. The presentation will cover how to:

• Make simulation models accurate under different conditions and countries. This can be done using realistic camera/radar/lidar modeling approaches, IAV competencies on perception modeling, virtual world creation, and generation of road networks and realistic assets for country-based validation.
• Create realistic scenes and scenarios from real-world test—including data augmentation, data analysis, AI-based generation of 3D models—both manually and programmatically. The outlook is to include multi-agent simulation.
• Combine all-new methods to create a complete solution package.