A Hybrid Approach to System Verification in Early Design for Complex Mechatronic Systems Based on Formal Functional Semantics
Yue Cao,Yusheng Liu,Xujia Qin
DOI: https://doi.org/10.1016/j.aei.2023.102201
IF: 8.8
2023-01-01
Advanced Engineering Informatics
Abstract:Model-based systems engineering (MBSE) has been adopted as a mainstream design methodology for complex mechatronic systems. According to this paradigm, structured models, typically represented in SysML are used to describe the system of interest from various aspects. Among these models, the system behavior model is fundamental for the subsequent design and hence should be verified against system functions in the early design stage. However, due to the lack of formal semantics of the involved SysML models and software-physical hybrid characteristics of the mechatronic systems, it is not a trivial task to apply existing formal verification methods to solve this problem. In this study, a novel approach relying on hybrid functional semantics is proposed. First, this verification problem is formally defined based on an in-depth analysis of the architectures and traceability between the involved SysML models. Then, a graphical modeling language to represent the hybrid functional semantics in SysML is defined so that the function and behavior models can be enhanced and explicitly correlated by the formal semantics. Finally, based on the semantically enhanced models, the model verification problem is formally re-defined as a semantics verification problem and further divided into two sub-problems, i.e., continuous and discrete semantics verification problems, which are automatically solved by leveraging existing verification techniques including functional-effect compatibility check and symbolic model checking. A mobile robot is illustrated to show the effectiveness of the proposed approach.