Digital Twin (DT) is one of the key enabling technologies for the Industry 4.0 revolution. It represents a comprehensive physical and virtual description of a product or a system. After a decade of development, DT has been applied to almost all industrial sectors covering different lifecycle phases. The benefit of DT is undoubted; however, it faces challenges when dealing with certain complex industrial systems. It requires integration of multiple relevant DTs corresponding to different system levels and lifecycle phases. It’s time to consider the next generation of DT.
The concept of Cognitive Digital Twin (CDT) has been recently proposed which reveals a promising evolution of the current DT paradigm towards a more intelligent, comprehensive, and full-lifecycle representation of complex systems.
CDT characteristics and definition
Although there is no wide spread consensus on the CDT definition, some common features can be extracted.
(1)DT-based: CDT is an extended or augmented version of DT. It contains at least the three basic elements of DT, including the physical entity, virtual representation, and the connections between the virtual and physical spaces. The difference is that a CDT may contain multiple DTs with unified semantics topology definitions. Particularly for a complex industrial system, its CDT should include digital models of its subsystems, and each of them has different status across the entire lifecycle.
- Cognition capability: As indicated in its name, a CDT should have certain human-like cognition capabilities, such as attention, perception, comprehension, memory, reasoning, prediction, decision-making, problem-solving, reaction and soon. Thus, a CDT is defined to recognise complex and unpredicted behaviours with optimisation strategies dynamically.
- Full life cycle management: A CDT should consist of digital models covering different phases across the entire life cycle of the system, including beginning-of-life(BOL, e.g.design, building, testing),middle-of-life (MOL, e.g.operating, usage, maintenance) and end-of-life (EOL, e.g.disassembly, recycling, re-manufacturing). It should also be capable of integrating and analysing all available data, information and knowledge from different life cycle phases thus to support a fore mentioned cognitive activities.
- Autonomy capability: A CDT should conduct autonomous activities without human assistance or minimum level of human intervention. This capability is partially over lapped with and empowered by the cognition capabilities of a CDT. For example, based on the perception and prediction results, a CDT can autonomously make decisions and react for design, production or operations adaptively.
- Continuous evolving: A CDT should be able to evolve along the entire system There exists three levels of evolving. First, for a single digital model, it updates itself according to the change of relevant data, information from the physical system; second, due to the interactions among different digital models contained in the same life cycle phase, each model evolves dynamically according to the impact of other models; third, due to the feedback from other life cycle phases, the previous two situations may happen simultaneously, even new models and components will be added.
More details about the concept of CDT can be found in the recently published paper available to download at the link https://doi.org/10.1080/00207543.2021.2014591, where the following CDT definition has been proposed:
“Cognitive Digital Twin(CDT) is a digital representation of a physical system that is augmentedwithcertaincognitivecapabilitiesandsupporttoexecuteautonomousactivities;comprisesasetofsemanticallyinterlinkeddigitalmodelsrelatedtodifferentlifecyclephasesofthephysical system including its subsystems and components; and evolves continuously with the physical system across the entire lifecycle.”
CDT enabling technologies and standardization
The main enabling technologies for CDT development and implementation directly correlate with the CDT characteristics, including Semantic Technologies and Industrial Ontologies, Model Based Systems Engineering (MBSE), Product Lifecycle Management (PLM) etc.
The couplings of these technologies is another challenging task of CDT applications which requires the support of advanced digital infrastructures and relevant management technologies. From Cyber-Physical System (CPS) perspective, it is critical to interconnect subsystems across different domains and lifecycle phases using adapters, brokers and other types of middleware mechanisms.
From another perspective, standardization is fundamental to achieve interoperability of data and tool for constructing CDT. The formal standardisation of DTs is the basis for enabling the CDT paradigm. Several Standards Developing Organisations (SDOs) are developing DT standards such as the International Organization for Standardization (ISO), World Wide Web Consortium – Web of Things (W3C WoT), the Industrial Internet Consortium (IIC) and the Plattform Industrie 4.0, etc. Despite the lack of a universal DT standard, some existing standards and protocols can be considered as substitute. For example, Plattform Industrie 4.0 provides the Asset Administration Shell (AAS) as a part of the RAMI4.0. ETSI Industry Specification Group (ISG) proposes the Next Generation Service Interfaces-Linked Data (NGSI-LD) APIs. The W3C WoT working group proposes the WoT Thing Description (WoT TD) which is an official W3C Recommendation. The recent published ISO 23247 standard defines the principles and requirements for developing DTs in manufacturing domain, and provides a framework to support the creation of DTs of observable manufacturing elements including personnel, equipment, materials, manufacturing processes, facilities, environment, products, and supporting documents.
As an emerging concept, CDT is still in an early stage of its development. There are many challenges to be resolved in order to fully realize its vision. This will depend on the advancement of the aforementioned enabling technologies and standardization efforts in the future.