Goal modeling for collaborative groups of cyber-physical systems with GRL: reflections on applicability and limitations based on two studies conducted in industry

In the future, cyber-physical systems (CPS) will increasingly need to have the capability to collaborate with other CPS during runtime to fulfill an overall purpose. This is, for instance, the case for autonomous driving where individual vehicles form a platoon in order to reduce energy consumption and improve the overall traffic flow. When engineering such CPS, it is important to understand the goals the individual systems should fulfill as well as the goals of collaborating groups the individual CPS are potentially part of during operation. In this context, the use of goal modeling techniques seems promising to systematically identify, document, and analyze the scope and context of individual CPS and groups of collaborating CPS in terms of goals and goal relationships. Even though many approaches for goal modeling have been proposed, only a few of them already found their way to industry - such as GRL (goal-oriented requirements language), which is also the essential part of the ITU Z.150/151 standard. In this paper, we report on empirical insights gained from applying the GRL for goal modeling of collaborative groups of CPS in the application fields 'autonomous driving' and 'collaborative transport robots'. We explore different challenges that originate from the very nature of dynamical collaboration of CPS and report on our findings from applying GRL in this context. In order to facilitate goal modeling for individual CPS and collaborative groups thereof, we also propose extensions to GRL addressing specific limitations of GRL identified in our studies.