Timing consistency checking for UML/MARTE behavioral models

Abstract

UML/MARTE model-driven development approaches are gaining attention in developing real-time embedded software (RTES). UML behavioral models with MARTE annotations are used to describe timing behaviors and timing characteristics of RTES. Particularly, state machine, sequence, and timing diagrams with MARTE annotations are appropriate to understand and analyze timing behaviors of RTES. However, to guarantee software correctness and safety, timing inconsistencies in UML/MARTE should be identified in the design phase of RTES. UML/MARTE timing inconsistencies are related to modeling errors and can be hazards throughout the lifecycle of RTES. We propose a systematic approach to check timing consistency of state machine, sequence, and timing diagrams with MARTE annotations for RTES. First, we present how state machine, sequence, and timing diagrams with MARTE annotations specify the behaviors of RTES. To overcome informal semantics of UML/MARTE models, we provide formal definitions of state machine, sequence, and timing diagrams with MARTE annotations. Second, we present the timing consistency checking approach that consists of a rule-based and a model checking-based timing consistency checking. In the rule-based timing consistency checking, we validate well formedness of UML/MARTE behavioral models in timing aspects. In the model checking-based timing consistency checking, we verify whether timing behaviors of sequence and timing diagrams with MARTE annotations are consistent with the timing behaviors of state machine diagrams with MARTE annotations. We support an automated timing consistency checking tool UML/MARTE timing Consistency Analyzer for a seamless approach. We demonstrate the effectiveness and the practicality of the proposed approach by two case studies using cruise control system software and guidance and control unit software