Reliability assessment of environment perception sensors

Sensors play a crucial role in the development of digitized and interconnected products which interact with the environment. They provide real time data for the automation and optimization of complex tasks that before have not been able or had to be performed solely by humans. Currently one of the most important examples in this context is the usage of sensors such as LIDAR, RADAR and cameras for advanced driver assistance systems (ADAS). Ultimately the aim is to enable fully automated driving in the near future.

When relying on environmental perception of sensors in safety critical applications, it is very important to precisely assess and predict the sensor’s accuracy and reliability. Depending on the level of automation, failure of the full sensor set in a specific application could lead to serious consequences. Therefore, before introducing new systems, it has to be proven that the associated risk is acceptably low.

This project is a collaborative project of the ERA Group and the AUDI AG, INI.TUM with the aim of developing a probabilistic framework that allows to assess the reliability of perception sensors in the context of automated vehicles.