UML4IoT—A UML-based approach to exploit IoT in cyber-physical manufacturing systems
Introduction
Manufacturing systems independent of their nature address the challenge of satisfying product customization needs. Customers are expecting to have products that will address their specific needs and will be comparable in cost to mass-produced ones [1]. Discrete process control systems, such as assembly systems [2], or batch process control systems [3] should gradually be transformed to highly adaptive and resource-efficient systems able to address the always increasing needs of product customization [4]. The industry has to address many challenges in order to successfully switch to this level of flexibility and retain its leading position in economy [5]. Multidisciplinary areas such as mechatronics and cyber-physical systems (CPS) as well as IT technologies such as Internet-of-Things (IoT) and cloud computing are playing a leading role in this industrial revolution, which is known as the fourth industrial revolution or Industry 4.0 [6].
The traditional approach for the development of manufacturing systems considers: (a) the system as a composition of the physical plant, the network of computation nodes and the computational processes required to monitor and control the physical ones, and, (b) the development of each one of these three discipline parts independent of the others by using their own specific frameworks, methods and tools. This approach is unable to address the demand for synergetic mechatronic dependability predictions [7] and is considered inadequate to address the increased requirements for flexibility and evolvability of today’s systems [8], [9]. It does not force an actual cooperation in the development of the three discipline parts, and it commonly leads to non-optimal solutions with high coupling between the main constituent components of the CPS, i.e., the physical plant, the network of computation nodes and the computational processes. Model Integrated Mechatronics [9] enhanced with the 3 + 1 SysML-view model [10] addresses this challenge by considering the system as a composition of well-defined reusable mechatronic components, i.e., cyber-physical components. A cyber-physical component is the tight integration of the physical world unit, i.e., the plant mechanical unit, with the cyber part that is required to transform it to a highly cohesive component that offers, through well-defined interface, higher level functionality, compared to the one offered by the physical unit. In this way, computing and communication capabilities have been embedded in the physical component transforming it to cyber-physical component, such as the one used in energy systems [11]. This approach greatly reduces the coupling between the system components compared to the traditional one and has already found the road to production in the context of Industry 4.0, e.g., FESTO [12] and factory-of-things [13]. The interface of the cyber-physical component is composed of physical, cyber-physical and cyber ports through which it is integrated with other components so as to effectively collaborate with these to accomplish the higher level of behavior that is required at the sub system or system level. The integration process of the constituent components of cyber-physical systems is a great challenge since it directly affects quality properties of the system such as maintainability, configurability, extensibility, adaptability, flexibility and evolvability [2].
Technologies such as the Internet of Things (IoT), Cloud computing, Service Oriented Architectures (SOA) and mobile computing, if successfully adapted to the industrial automation domain may address challenges in modern manufacturing. Web standards such as SOAP and WSDL have already been adopted by research groups in the industrial automation domain and several approaches have been described to exploit their benefits, as for example [14], [15], [16] to mention a few. SOA based products have already been introduced in the industrial systems market in the context of Industry 4.0. For example, TwinCAT from Beckhoff combines IEC 61131-3-based SOA services with OPC UA interoperability [17]. However, SOAP and WSDL have been proved too heavyweight compared to the recent IoT application layer protocols such as CoAP and MQTT. Furthermore, IoT is well-aligned with the architecture of a manufacturing enterprise and as authors argue in [5] it is able to provide “vital solutions to planning, scheduling, and controlling of manufacturing systems at all levels.” IoT brings great opportunities in achieving better system performances in globalized and distributed environments. However, the IoT is in infant stage in manufacturing and there is a demand for “research, development and standardization of enabling technologies for safe, reliable, and effective communication and decision-making” [5]. The developer when dealing with the actual deployment of IoT apps, especially in complex and sensitive scenarios, such as those in the manufacturing industry, has a tedious task, requiring expertise in disparate fields, as also claimed in [18]. Even, the creation of a simple IoT prototype requires knowledge on different heterogeneous tools, programming platforms and protocols, that results in increased time, effort and development costs. There is a need for platforms to provide among others information integration [19]. Thus, methodologies and approaches that will streamline the integration of manufacturing components using IoT will play a leading role in the effective exploitation of IoT in the domain of cyber-physical manufacturing systems and this is a challenge for academia and industry.
The approach presented in this paper, namely UML4IoT, effectively integrates trends in cyber-physical systems and IoT and describes a framework to address challenges introduced by the use of IoT in the development process of manufacturing systems. The IoT technologies that have been adopted in this work, i.e., the Open Mobile Alliance (OMA) LWM2M application protocol [20] and the smart objects [21] defined by the Internet Protocol for Smart Objects (IPSO) Alliance, focus on modeling the exposed interface of simple smart objects. They are not able to address the modeling needs of complex components of manufacturing systems, which are effectively addressed using the system modeling language SysML in collaboration with UML, as for example in [22]. Cyber-physical components modeled with SysML and UML appear to have the conventional object-oriented (OO) API, which is expressed as a set of classes with their associated methods and leads to a high coupling between the components. LWM2M and IPSO smart objects may be properly used to develop a layer on top of the conventional OO API to transform the cyber-physical component to an IoT-compliant component, i.e., to an Industrial Automation Thing. This layer is referred in this work as the IoTwrapper; it is the layer that transforms the OO API to a RESTful one.
Our approach automates the generation process of the IoTwrapper for new cyber-physical components but also for legacy ones to exploit the IoT connectivity. Two alternatives are presented and discussed. The first one is based on the UML design specification of the cyber part of the cyber-physical component; the second one is based on the source code, in case that a higher level design specification such as the UML one is not available. Java is used as a case study although other languages, for example the IEC 61131, can also be considered. The presented approach integrates modeling techniques required for the specification of complex cyber-physical components, i.e., SysML and UML, with IoT technologies, i.e., OMA LWM2M and IPSO objects, for the construction of the Industrial Automation Thing, which is considered as the key building block of the modern IoT manufacturing environment. Extensions are proposed to the UML modeling notation so as to enable an automatic generation of the IoTwrapper of the component that transforms it to an Industrial Automation Thing. The IoT-compliant interfaces, if properly used at the system or subsystem integration level may lead to on demand system configurations that address specific customer needs in a cost effective way. The presented approach is in line with the perspective presented in Ref. [23] based on which many of the smart components associated with the IoT will be mechatronic in nature, which imposes the need for significant changes to the way mechatronic, and related, systems are designed and configured.
The low level control of the physical unit is embedded in the cyber-physical components so it is not affected by the use of the IoT. However, performance is always an issue in manufacturing systems even in the interaction among cyber-physical components. This is why a performance analysis of the proposed layer is of great interest to the industrial engineer in order to able to estimate the performance overhead introduced by the IoTwrapper in the interaction of the system’s components. The main contributions of this paper are: (a) a novel approach to address the problem of effective integration of cyber-physical components in the modern IoT manufacturing domain, (b) the definition of a UML profile for the exploitation of IoT in this domain, (c) the automation of the generation process of the IoTwrapper of the cyber-physical component, and (d) a lightweight flexible prototype implementation of the OMA LWM2M protocol based on metaprogramming. The proposed UML profile contains basic key constructs independent of the application domain so as to be general enough to be applicable in other application domains.
The remainder of this paper is structured as follows. In the next section, the approach described in this paper as well as the example system used as case study are briefly described. In Section 3, related work is presented. The UML profile for IoT and its exploitation to automate the generation process of the IoTwrapper of the cyber-physical component is presented in Section 4. The process for the automation of the generation of the Industrial Automation Thing is described in Section 5. Performance evaluation of the IoTwrapper is given in Section 6 and the paper is concluded in the last section.
Section snippets
The myLiqueur production system
The liqueur plant system used as case study in [24] was adopted as base for the definition of the myLiqueur production system, which exploits IoT to allow end users to produce custom types of liqueur. Production parameters that define the specific type of liqueur could be defined by the end user through the myLiqueur App. The myLiqueur production system is composed of the following cyber-physical components, as shown in Fig. 1: smartSilo1, smartSilo2, smartSilo3, smartSilo4 and smartPipe. Each
Related work
Cyber-physical systems, which are considered as the orchestration of the computational and physical processes that constitute the manufacturing system [29], play an important role towards Industry 4.0. The great impact of CPSs in manufacturing based on a number of explorative case studies is examined in Ref. [30]. Authors argue that CPSs are transforming the service business in manufacturing and offer new opportunities for business innovation. Real-time requirements on manufacturing systems as
UML4IoT—a LWM2M-based UML profile for IoT
For the effective integration of the cyber-physical component to the modern IoT manufacturing environment, its cyber part not only has to offer the services provided by the component to the environment but also has to support the management of the component, its monitoring and configuration, as well as its maintenance and repair. Interoperability is also a key requirement for this integration. To address these requirements we have adopted the OMA LWM2M. OMA is developing the LWM2M standard to
Automating the generation of the IoTwrapper
In the case that the UML design specification is not available for the cyber part of the cyber-physical component, an alternative approach based on the source code of the cyber part was defined for the generation of the IoTwrapper. The basic idea is to annotate the source code of the cyber part of the component using specific annotations defined for this reason. The application of the UML4IoT in the case that the cyber part is developed using the IEC 61131 function block model, which is widely
Measurements and evaluation
To evaluate the timing behavior of the IPSO-compliant cyber-physical component and the overhead introduced by the IoTwrapper, a number of measurements have been performed using as test bed the prototype implementation of the liqueur production system. Two implementations of the IoTwrapper are used in the measurements.
In the first deployment scenario the wrapper has been developed using the leshan implementation of the OMA LWM2M. In this case the IoTwrapper, i.e., the leshan wrapper, was
Conclusions
IoT is transforming the way that modern manufacturing systems will be developed and will operate. For example, the introduction of the REST architectural paradigm greatly influences the development process. The adoption of IoT imposes a paradigm shift for the automation system developer and complicates the development process. Effective approaches are required to handle the complexity introduced by such a transition. Moreover, there is a need for legacy manufacturing components to be integrated
Acknowledgments
Authors would like to thank the leshan development team and more specifically Simon Bernard and J.F. Schloman for the support on using leshan. Authors would also like to thank the anonymous reviewers for the constructive comments that resulted in an improved version of the paper.
References (70)
- et al.
Formal computer-aided product family architecture design for mass customization
Comput. Ind.
(2015) - et al.
Current status and advancement of cyber-physical systems in manufacturing
J. Manuf. Syst.
(2015) On conceptual design of intelligent mechatronic systems
Mechatronics
(2003)SmartFactory—towards a factory-of-Things
Annu. Rev. Control
(2010)A cyber-physical system-based approach for industrial automation systems
Comput. Ind.
(2015)- et al.
The Internet of Things—the future or the end of mechatronics
Mechatronics
(2015) - et al.
The impact of cyber-physical systems on industrial services in manufacturing
- et al.
State-of-the-art, challenges, and open issues in the integration of Internet of things and cloud computing
J. Network Comput. Appl.
(2016) - et al.
A cyber-physical systems architecture for industry 4.0-based manufacturing systems
Manuf. Lett.
(2015) Machine availability monitoring and machining process planning towards cloud manufacturing
CIRP J. Manuf. Sci. Technol.
(2013)
A domain specific language for high-Level process control programming in smart buildings
Domain-specific languages: a systematic mapping study
Inf. Software Technol.
Extracting reusable design decisions for UML-based domain-specific languages: a multi-method study
J. Syst. Software
An open distributed architecture for flexible hybrid assembly systems: a model driven engineering approach
J. Adv. Manuf. Technol.
An IEC 61499 based approach for distributed batch process control
Internet of things for enterprise systems of modern manufacturing
IEEE Trans. Ind. Inf.
Recommendations for implementing the strategic initiative industrie 4.0
Final Report of the Industry 4.0 Working Group
Dependability Modelling Under Un-certainty: An Imprecise Probabilistic Approach
Model integrated mechatronics—towards a new paradigm in the development of manufacturing systems
IEEE Trans. Ind. Inf.
The 3 + 1 SysML view-model in model integrated mechatronics
J. Software Eng. Appl. (JSEA)
Innovative smart grid technologies (ISGT latin america)
2011 IEEE PES Conference, 19–21 October
Design and implementation of a service-oriented architecture for the optimization of industrial applications
IEEE Trans. Ind. Inf.
A real-time service-oriented architecture for industrial automation
IEEE Trans. Ind. Inf.
Semantics-based composition of factory automation processes encapsulated by web services
IEEE Trans. Ind. Inf.
Industrial application development exploiting IoT vision and model driven programming
Intelligence in Next Generation Networks (ICIN), 2015 18th International Conference
Cloud computing for industrial automation systems—a comprehensive overview
On the implementation of industrial automation systems based on PLC
IEEE Trans. Autom. Sci. Eng.
Cited by (148)
CHESSIoT: A model-driven approach for engineering multi-layered IoT systems
2024, Journal of Computer LanguagesA formal approach to specify and verify Internet of Things architecture
2023, Internet of Things (Netherlands)Correctness of IoT-based systems: From a DSL to a mechanised analysis
2023, Journal of Computer LanguagesFormal verification of IoT applications using rewriting logic: An MDE-based approach
2022, Science of Computer ProgrammingAutomatic generation of Web of Things servients using Thing Descriptions
2024, Personal and Ubiquitous ComputingWHAT HINDERS INTERNET OF THINGS (IOT) ADOPTION IN THE CHINESE CONSTRUCTION INDUSTRY: A MIXED-METHOD
2024, Journal of Civil Engineering and Management