Abstract Digital twins (DTs) are proliferating in a multitude of domains, including agriculture, automotive, avionics, logistics, manufacturing, medicine, smart homes, etc. As domain experts and software experts both have to contribute to the engineering of effective DTs, several model-driven engineering (MDE) approaches have been recently proposed to ease the design, development, and operation of DTs. However, the diversity of domains in which MDE is currently applied to DTs, as well as the diverse landscape of DTs and MDE applications to DTs, makes it challenging for researchers and practitioners to get an overview of what techniques and artifacts are already applied in this context. In this paper, we shed light on the aforementioned aspects by performing a systematic mapping study on the application of MDE automation techniques, i.e., model-to-model transformation, code generation, and model interpretation, in the context of DTs as well as on the characteristics of DTs including the twinned systems to which these techniques are applied in different domains. We systematically retrieved a set of 189 unique publications, of which 66 were selected for further investigation in this paper. Our results indicate that the distribution of employed MDE techniques (136 applications of automation techniques) is balanced between the different techniques, but there are significant variations for different DT types. With respect to the different domains, we found that even though applications are available in many domains, a small number of domains currently dominate applications of MDE to DTs, i.e., more than half of included papers are in the manufacturing and transportation domains.

For continuously checking and updating the virtual representation of a real system during operation, the continuous sensing and interpretation of raw sensor data is a must. The challenge is to bundle sensor value streams (e.g., from IoT networks) and aggregate them to a higher logical state level to enable process-oriented viewpoints and to handle uncertainties about sensor measurements and state realization precision. To address these uncertainties, so-called “tolerance ranges” must be defined in which logical states are detected during operation with acceptable deviations. Specifying such tolerance ranges manually is a time-consuming, error-prone task and often not feasible due to the huge associated value search space. To tackle this challenge, the problem is turned into an optimization problem in this paper. For this purpose, we present a framework based on meta-heuristic search that enables the automatic configuration of tolerance ranges based on available execution traces of multiple sensor value streams. An exploratory study evaluates the approach. For this purpose, we implemented a lab-sized demonstrator of a five-axis grip arm robot, which we continuously monitored during operation in a simulated environment. The evaluation shows the advantage of using meta-heuristic optimizers such as Harmony Search or Genetic Algorithm to identify stable tolerance ranges automatically for state detection at runtime.Note to Practitioners—Monitoring sensor values streams is nowadays a frequently employed technique in many automation domains. However, combining and mapping single value streams to higher-level state-based representations such as state machines or other design-time related models is a major challenge due to measurement and realization precision uncertainties. Thus, simply mapping monitored raw data to these design descriptions can lead to falsely identified or missed states. To improve this situation, we present an approach that provides a mechanism to continuously analyze data streams during operation by automatically finding appropriate tolerance ranges to detect realized system states. The approach uses a small set of annotated execution traces and meta-heuristic searchers to derive optimal tolerance ranges, which provide high correctness and completeness of the identified system states. This approach represents the basis for building a “vertical bridge” from the operation technology layer considering pure sensor data streams to the IT layer where state-based process views are provided to perform monitoring and analytics, e.g., by using process mining.

Model-Based Systems Engineering (MBSE) has emerged with great potential to fulfill the non-linearly rising demand in interdisciplinary engineering, e.g., product development. However, the variety and complexity of MBSE tools pose difficulties in particular industrial applications. This paper tries to serve as a guideline to find the ideal tool for a specific industrial application as well as to highlight the key criteria that an industry might consider. For this purpose, we propose a logical framework for MBSE tool selection, which is based on market research, the approaches of Quality Function Deployment (QFD), and decision matrix. As customers are at the center of any product, accordingly the needs of MBSE tool users are addressed within this research as the fundamental starting point. Market research and extensive discussions with MBSE tool vendors and academia show the current situation of MBSE tools. To compare the performance of the considered tools, a set of user needs is defined. QFD is performed to analyze the user needs with respect to evaluable technical properties. Subsequently each tool performance is assessed using a decision matrix. Through this process, a well-defined functional structure of MBSE tools is sketched, and in order to identify the properties of an ideal tool, all the attributes of different MBSE tools are mapped to a common platform. For the purpose of evaluation, we apply our proposed logical framework to select an exemplary MBSE tool for interdisciplinary application.

As technologies such as the Internet of Things (IoT) and Cyber-Physical Systems (CPS) are becoming ubiquitous, systems adopting these technologies are getting increasingly complex. Digital Twins (DTs) provide comprehensive views on such systems, the data they generate during runtime, as well as their usage and evolution over time. Setting up the required infrastructure to run a Digital Twin is still an ambitious task that involves significant upfront efforts from domain experts, although existing knowledge about the systems, such as engineering models, may be already available for reuse. To address this issue, we present AML4DT, a model-driven framework supporting the development and maintenance of Digital Twin infrastructures by employing AutomationML (AML) models. We automatically establish a connection between systems and their DTs based on dedicated DT models. These DT models are automatically derived from existing AutomationML models, which are produced in the engineering phases of a system. Additionally, to alleviate the maintenance of the DTs, AML4DT facilitates the synchronization of the AutomationML models with the DT infrastructure for several evolution cases. A case study shows the benefits of developing and maintaining DTs based on AutomationML models using the proposed AML4DT framework. For this particular study, the effort of performing the required tasks could be reduced by about 50%.

In the area of Cyber-Physical Systems (CPS), the degree of complexity continuously increases mainly due to new key-enabling technologies supporting those systems. One way to deal with this increasing complexity is to create a digital representation of such systems, a so-called Digital Twin (DT), which virtually acts in parallel ideally across the entire life-cycle of a CPS. For this purpose, the DT uses simulated or real-time data to mimic operations, control, and may modify the CPS's behaviour at runtime. However, building such DTs from scratch is not trivial, mainly due to the integration needed to deal with heterogeneous systems residing in different technological spaces. In order to tackle this challenge, Model-Driven Engineering (MDE) allows to logically model a CPS with its physical components. Usually, in MDE such "logical models" are created at design time which keep them detached from the deployed system during runtime. Instead of building bilateral solutions between each runtime environment and every engineering tool, a dedicated integration layer is needed which can deal with both, design and runtime aspects. Therefore, we present a reference architecture that allows on the one side to query data from model repositories to enrich the running system with design-time knowledge, and on the other side, to be able to reasoning about system states at runtime in design-time models. We introduce a model repository query and management engine as mediator and show its feasibility by a demonstration case.

The OMG standard Systems Modeling Language (SysML) has been on the market for about thirteen years. This standard is an extended subset of UML providing a graphical modeling language for designing complex systems by considering software as well as hardware parts. Over the period of thirteen years, many publications have covered various aspects of SysML in different research fields. The aim of this paper is to conduct a systematic mapping study about SysML to identify the different categories of papers, (i) to get an overview of existing research topics and groups, (ii) to identify whether there are any publication trends, and (iii) to uncover possible missing links. We followed the guidelines for conducting a systematic mapping study by Petersen et al. (Inf Softw Technol 64:1–18, 2015 ) to analyze SysML publications from 2005 to 2017. Our analysis revealed the following main findings: (i) there is a growing scientific interest in SysML in the last years particularly in the research field of Software Engineering, (ii) SysML is mostly used in the design or validation phase, rather than in the implementation phase, (iii) the most commonly used diagram types are the SysML-specific requirement diagram, parametric diagram, and block diagram, together with the activity diagram and state machine diagram known from UML, (iv) SysML is a specific UML profile mostly used in systems engineering; however, the language has to be customized to accommodate domain-specific aspects, (v) related to collaborations for SysML research over the world, there are more individual research groups than large international networks. This study provides a solid basis for classifying existing approaches for SysML. Researchers can use our results (i) for identifying open research issues, (ii) for a better understanding of the state of the art, and (iii) as a reference for finding specific approaches about SysML.

Nowadays, software-as well as hardware systems produce log files that enable a continuous monitoring of the system during its execution. Unfortunately, such text-based log traces are very long and difficult to read, and therefore, reasoning and analyzing runtime behavior is not straightforward. However, dealing with log traces is especially needed in cases, where (i) the execution of the system did not perform as intended, (ii) the process flow is unknown because there are no records, and/or (iii) the design models do not correspond to its real-world counterpart. These facts cause that log data has to be prepared in a more user-friendly way (e.g., in form of graphical representations) and algorithms are needed for automatically monitoring the system's operation, and for tracking the system components interaction patterns. For this purpose we present an approach for transforming raw sensor data logs to a UML or SysML sequence diagram in order to provide a graphical representation for tracking log traces in a time-ordered manner. Based on this sequence diagram, we automatically identify interaction models in order to analyze the runtime behavior of system components. We implement this approach as prototypical plug-in in the modeling tool Enterprise Architect and evaluate it by an example of a self-driving car.