面向信息物理融合系统的异构模型转换方法

Heterogeneous model translation method for the cyber physical system

作为一种深度网络化嵌入式系统,信息物理融合系统(Cyber-Physical System,CPS)具有联合动态性、系统组成异质性、计算过程与物理过程的多尺度融合性等典型特征,使得基于单一模型的传统嵌入式系统模型构建和仿真验证方法面临新的挑战.在CPS系统设计开发与仿真验证时,往往需要组合使用多种设计模型或建模语言以描述信息域与物理域实体特点.针对CPS计算物理深度融合的问题,在分析基于计算模型和物理模型进行CPS系统协同开发需求和可行性的基础上,以统一建模语言(Unified Modeling Language,UML)模型和仿真(Simulink)模型分别作为计算过程和物理过程典型建模方式,研究了连续时间和离散事件模型间的结构映射和行为映射,提出了一种Simulink模型与UML类图和活动图之间的转换方法,并通过ATL(ATLAS Transformation Language)转换规则实现了技术验证.

As a kind of deeply networked embedded system,the Cyber-Physical system（CPS） has been characterized by the joint dynamics among computation,and heterogeneous composition,multi-scale fusion between computing process and physical process etc.It is inadequate and difficult for the single model design methods for the traditional embedded computing system to adapt to the requirements of CPS modeling and simulation.In the process of design,development and simulation,the CPS often needs to combine multiple design models or to use the modeling language to describe the entity of the computing domain and physical domain.Towards the computing-physical depth fusion of in the CPS,based on the analysis of the demand and feasibility of the computing model and physical model collaborative development in the CPS,this paper use the unified modeling language（UML） model and Simulink model as the computing process and physical process modeling method separately,studies the structure mapping and behavior mapping between continuous time model and discrete event model. An approach to the transformation between the Simulink model and the UML class diagram and activity diagram is presented and verified by ATLAS transformation language（ATL） technology to realize the conversion rules.