The Quest for Infrastructures and Engineering Methods Enabling Highly Dynamic Autonomous Systems

The trends for connected and automated vehicles and their related technologies are radically changing the automotive industry: from engineering automotive systems, through their production and down to their use. All of these phases will be characterized by a high level of dynamism: the driving behavior of automated cars and production processes in a factory will likewise be controlled by learning and adaptive systems that evolve over time to deal with quickly changing tasks and environments. Automated cars on the roads and robots in factories are both highly mobile and need to collaborate in dynamic neighborhoods. What are the ICT infrastructures and engineering methods that are required to build such highly dynamic but at the same time trustworthy autonomous systems? In this position paper we argue that in the context of infrastructure, a “Decimeter Internet” will be needed that offers low-enough communication latencies, adequately accurate localization plus update rates equal to the task of dependably controlling mobile cars and robots at decimeter level. With respect to engineering methods, we argue the strong need for approaches, which will ensure the dependability of learning and adaptive systems that change their behavior during runtime.