Safety engineering for autonomous vehicles

In safety engineering for non-autonomous vehicles, it is generally assumed that safety is achieved if the vehicle appropriately follows certain control commands from humans such as steering or acceleration commands. This fundamental assumption becomes problematic if we consider autonomous vehicles that decide on their own which behavior is most reasonable in which situation. Safety criticality extends to the decision-making process and the related perception of the environment. These, however, are so complex that they require the application of concepts for intelligence that do not harmonize with traditional safety engineering. In this paper, we investigate these problems and propose a solution.