Modeling collaborations with dynamic structural adaptation in mechatronic UML

Abstract

The next generation of advanced mechatronic systems is expected to behave more intelligently than today's systems. These systems are expected to enhance their functionality and improve their performance by building communities of autonomous agents which exploit local and global networking. Such mechatronic systems will therefore include complex coordination protocols which require execution in real-time and reconfiguration of the locally employed control algorithms at runtime to adjust their behavior to the changing system goals leading to self-adaptation. In this paper we will present an extension of the MECHATRONIC UML approach which will enable us to model collaborations between components which include structural adaptation and multi-ports. Besides the modeling of complex collaborations and the rules to join and leave these collaborations via ports and multi-ports, we propose to employ hierarchical state machines with a dynamic number of submachines to model the behavior of the multi-ports. For the collaborations this involves the related protocols, while for the components we have to refine this behavior to ensure a proper synchronization with other parts of the component behavior.

Modeling Collaborations in Adaptive Systems of Systems

“…The rigSoaML modeling approach described in [30] extends the collaboration modeling for service-oriented systems by means of a verification concepts of joint system interactions. In the context of embedded systems, the Mechatronic UML (mUML) approach [100] facilitates the modeling, verification and simulation of collaboration behavior. None of these approaches above facilitate the explicit modeling of collaborations in adaptive SoS as done in this thesis.…”

“…Moreover, the architecture is predefined and extensions or changes in the proposed approach are not discussed. Also previous work [53,100] proposes the Mechatronic UML approach that extends the UML to specify and generate a hierarchical architecture. The approach addresses distinct feedback loops for reconfiguration and planning purposes.…”

“…For example, Hirsch et al [100] use graph transformation rules for varying the system architecture for coordination purposes during collaboration in mechatronic systems. Furthermore, Vogel et al [174] use triple graph grammar rules for the runtime deployment of components in an adaptive web shop example.…”

“…Because of an underlying toolchain described in [52], mUML models can be analyzed and simulated. Another extension of mUML introduces a collaboration concept [100], which already has a notion of collaboration roles. This concept enables an interaction between components and extends the static reconfiguration capabilities to dynamic structural adaptations at runtime.…”

Graph based verification of software evolution requirements

“…Because of its popularity and extensibility, variations of UML models are proposed in the literature for modeling special types of systems, such as mechatronic [69] and real-time systems [16]. These models are extended with formal verification tools that specialize on verifying runtime reconfiguration of these systems.…”