Virtual Validation of Warning-Based ADAS Features
Divya T, Ashok Leyland
Gururaaj P, Ashok Leyland
Advanced driver-assistance systems (ADAS) are becoming crucial for enhancing the safety of road users and vehicles. Developing ADAS features requires the development of sophisticated algorithms, involving sensors and control systems to perceive the environment, make decisions, and execute driving maneuvers. Validating ADAS features, such as lane departure warning (LDW) and forward collision warning (FCW), is essential for improving vehicle safety, especially on Indian roads. To ensure these features are robust, extensive validation is required. Virtual validation offers a scalable and efficient solution for this purpose. In this session, we talk about some of the key components of algorithm and validation:
- Automating data annotation for vision perception algorithm development
- Accurately creating realistic scenes and scenarios for virtual validation including custom 3D assets
- Modeling virtual sensors (cameras) and vehicles (trucks) to reflect real-world conditions
- Integrating ADAS algorithms in a virtual environment to evaluate performance and reliability in various scenarios
- Creating custom metrics to assess the performance of FCW and LDW features
Published: 22 Dec 2024