Evaluation of Novel Safety Concepts for Automotive Perception in Real-World Environments

Safety is one of the most important topics regarding Automated-Driving in the Automotive Domain. In the last years, LiDAR, Radar and Vision Cameras became the most relevant Environmental Perception Sensors for enabling safe and robust Automated-Driving. All of these systems offers specific strengths and weaknesses in specific situations such as bright sunlight or heavy rain. Therefore, these systems requires specific Fail-Operational concepts to allow robust and safe driving in urban and rural environments.
In this publication, we are depicting Real-World evaluation of Safety Concepts and Fail-Operational functionality of a Sensor Fusion Platform that offers Radar, 3D Flash LiDAR and Vision Cameras. We verified our platform in specific driving situations such as driving from an urban parking environment into bright sunlight with the dynamic adaption of the confidence range that depicts reliable data.