Real-World vs. Simulation - Testing Camera-Based Functions for AD
To ensure that camera-based functions for autonomous driving systems are working properly, test cases must be relevant and must include scenarios that factor in different weather and lighting conditions such as night-time driving, snow, rain and fog. Using a combination of real camera data and simulated sensor camera data, numerous scenarios can be recreated in the lab to achieve an extensive and multifaceted test case study.
In this Learning Bits podcast, Gregor Sievers, Manager of Autonomous Driving Systems at dSPACE, shares different approaches that can be used in testing to mix and match simulation possibilities, including data replay, closed- loop simulation, fully virtual simulation, and incorporating pre-recorded videos.
Product Manager, Autonomous Driving Systems, dSPACE GmbH