In model-based systems engineering (MBSE) practice, traceability between the requirement and architecture models is crucial to ensure that the designed architectures can meet system requirements. Traceability can be defined as links from predefined requirements to the corresponding system architectures. Currently, the semantic heterogeneity of the requirement and system architecture models hinders semantic interoperability, causing difficulty in traceability management. To address this challenge, this paper proposes an ontology formalism and semantic rules for traceability management in MBSE. The proposed ontology formalism provides a unified specification that supports the semantic representation of requirement models, system architecture models, and their traceability links. Following the proposed ontology formalism, the semantic rules can effectively support both recognising the traceability links between the requirements at different levels of granularity and system architectures and identifying the requirements or system architectures affected by changes. A case study of automated driving systems is conducted to evaluate the completeness and correctness of the proposed ontology formalism and semantic rules for traceability management. The results demonstrate that the proposed formalism and semantic rules can significantly facilitate the establishment, maintenance, and utilisation of traceability links.

Embodied intelligence (EI) is the latest paradigm in the evolution of artificial intelligence (AI), which becomes a key enabler for advancing intelligent manufacturing in industrial scenarios. In recent years, the explosive growth of large language models (LLMs) has provided a critical pathway for realizing EI, rapidly spawning numerous groundbreaking applications. The widely application of industrial robots is the core characteristics of intelligent manufacturing, where the emergence of EI makes industrial robots becoming smarter and smarter with the capability of executing ever more complex tasks today. Consequently, research on LLMs for industrial EI has rapidly attracted attention from researchers worldwide in the past few years. To better understand the current situation in this field, this paper reviews the progress of LLMs towards industrial EI and discusses the corresponding research status and development trend from the perspectives of the capability breakthrough of single-modal LLMs, the fusion enhancement of multimodal large language models (MLLMs), and the embodiment realization of LLMs. This review aims to provide useful references for the further advancement of LLMs to support the development of industrial EI.

Model-Based Systems Engineering (MBSE) provides a formal and integrated framework for managing multidisciplinary information throughout the system design lifecycle. Within this framework, Measures of Effectiveness (MOEs) serve as the key quantitative metrics for evaluating whether a system design meets mission requirements. Current approaches for verifying MOE constraints in MBSE typically rely on mathematical relationships established through SysML parametric diagrams, with metrics analysis performed using built-in computational plug-ins or external specialized software. However, these approaches face challenges in unified representation of multitheory hybrid constraints, automated exploration for feasible solutions, and tool-level data consistency. This limits the reliability and completeness of verification results. This paper proposes a formal verification approach for MOE by integrating the multi-architecture modeling language KARMA with Satisfiability Modulo Theories (SMT). The approach provides a unified firstorder logic (FOL) representation of MOE-related metrics. It also automates the evaluation of SysML parametric diagrams through SMT solver, enabling constraint satisfiability checking and reverse derivation of optimal design solutions. A satellite electric power system (EPS) case study demonstrates the feasibility and effectiveness of the proposed approach. The results show that the approach ensures the consistency between system metrics and requirements, and improves automation and completeness of MOE verification within the MBSE framework.

With the increasing complexity of systems, ModelBased Systems Engineering (MBSE) has become the mainstream approach for managing structured information through digital models and enabling automated verification. However, enterprises have accumulated substantial engineering documents over decades, which preserve valuable design knowledge and domain expertise. In the context of digital engineering, digital artifacts such as documents and digital models constitute complementary elements that will coexist throughout the industry’s digital transformation. Transforming existing documents into MBSE models therefore represents a sustained capability for minimizing MBSE adoption cost while preserving organizational knowledge assets. Yet, automated document to model transformation faces significant challenges due to incompatible tool formats, diverse document structures, and knowledge scattered across unstructured content. This paper presents an agile knowledge reuse framework based on the Cognitive Digital Thread (CDT) concept. The framework employs Open Service for Lifecycle Collaboration (OSLC) compliant adapters for documents and modeling tools to unify content and operations as service interfaces. Mapping rules derived from document structures guide a rule engine to automatically extract document knowledge and generate system architecture models. A case study of landing gear system design demonstrates that the generated models remain consistent with their source documents, providing a practical solution for agile MBSE transformation based on existing knowledge assets.

With the advent of Industry 5.0, human-centric smart manufacturing is becoming a new paradigm for industrial transformation. Human-robot collaboration (HRC) is the hot topic of human-centric smart manufacturing. The emergence of large language model (LLM) provides significant opportunity for collaborative robot to promote the autonomous collaboration ability, which brings HRC into new era driven by embodied intelligence and more powerful robot. Therefore, a dynamic autonomous collaboration method inspired from looking-thinking-doing chain of human operators is proposed for human-like collaborative robot (HLCobot) in human-centric smart manufacturing based on multimodal large language model (MLLM), where perception-decision-execution coordination mechanism is constructed to appropriately distribute the abilities of MLLM in the dynamic operation chain of HRC. Firstly, a brain-inspired architecture with the integration of perception hub, decision hub, and execution hub is designed for dynamic autonomous collaboration. Secondly, the abilities of perception, decision, execution of HLCobot are realized by integrating MLLM, where the HLCobot can actively recognize the dynamic changes of HRC scenario by mimicking human operator and execute the correct motions to complete the necessary collaborative task autonomously. Additionally, a coordination mechanism among the agents of perception, decision, and execution is put forward to proceed the collaborative task smoothly. Finally, a case study of engine assembly is provided to demonstrate the effectiveness of the proposed method.

The rapid globalization of supply chains (SC) has opened vast prospects but also increased disruption risks and substantial uncertainties in supply chain systems-of-systems (SCSoSs). Supply chain reconfiguration (SCR) has emerged as a pivotal strategy for mitigating these risks. This paper proposes a multi-agent reinforcement learning-based resilience reconfiguration approach for SCSoSs to address the agile, stable, and spatio-temporal requirements of SCR under disruption risks. It begins by detailing the SCR issue involving suppliers, manufacturers, distributors, and consumers amid disruption risks and introduces three resilience strategies: filling, repairing, and recruiting. A three-phase model for calculating resilience and reconfiguration costs is then developed, grounded in the supply chain directed network (SCDN). Following this, the reconfiguration process is modeled as a partially observable Markov decision process (POMDP), with the state space representing SC elements and the action space including available strategies. The reward function balances resilience and costs considerations. Utilizing the multi-agent proximal policy optimization (MAPPO) technique, the method enables dynamic reconfiguration of SCSoSs, demonstrating its effectiveness through experimental simulations. The analysis also explores how different attributes affect reconfiguration outcomes. Results indicate that the MAPPO approach substantially enhances reconfiguration performance under disruption risks compared to other baselines, providing valuable insights for modern SC management.

Abstract The importance of domain knowledge and problem framing in design cognition is well established; the relationship between these factors, constraint-boundary understanding, model-derived information-acquisition behavior, and design performance in the context of the SIADM framework has not been examined extensively across different types of participants. In particular, there is insufficient empirical evidence on the role of these factors in a sequential information-acquisition decision-making task. This paper aims to address this by (1) adopting an existing SIADM model to make decisions when provided with information through a bounded-rational, myopic expected improvement decision rule, and (2) empirically assessing this model through a controlled experiment. The experiment measures an individual's domain knowledge using a Concept Inventory test, and the treatment variable is controlled by problem framing. Participants are given a design optimization challenge in two distinct approaches to address the research question: a Track Design Problem (TDP), presented as a domain-specific task, and a function-optimization problem (FOP). The result is interpreted in the context of the adopted SIADM framework and shows that the association between domain knowledge and domain-specific framing is associated with improved design outcomes, while the expected-improvement-rule is considered a descriptive approximation rather than a model of human decision-making.

Under the trend of human-centered intelligent manufacturing, the design of complex products is confronted not only with engineering challenges related to multidisciplinary integration, data complexity, and evolving requirements but also with the technical challenges arising from the paradigm shift in design processes due to autonomous design intelligence driven by AI technology. To investigate the emerging dimensions of human-centered design productivity, this paper proposes a cognition-decision bimodal synergy framework based on reinforcement learning for the generative design of complex products. This framework enables the automatic generation and optimization of design solutions by integrating design cognition and decision support knowledge into a unified knowledge triplet, which facilitates dynamic interactions between the agent and the environment. The framework consists of three primary components: cognitive and decision representations, state modeling of design solutions, and reinforcement learning model construction. To validate the feasibility and generalizability of the proposed framework, this study presents case studies involving architectural design, parametric design, and process design in the contexts of launch vehicles, autonomous mobile robots, and diesel engines. The proposed framework offers valuable insights for the future paradigm of intelligent complex product design and holds significant theoretical and practical implications for advancing the intelligent development of complex product design.

Ontology offers a formalized framework for knowledge representation and reasoning in complex systems, enabling the development of model-based systems engineering (MBSE). This study conducts a systematic bibliometric analysis of 566 papers on ontology in MBSE from 2008 to 2024, complemented by a qualitative literature review. The findings reveal a marked upward trajectory in research activities, characterized by smallscale collaboration networks predominantly led by core scholars. The research topics converge on several key topics related to the subject of ontology in MBSE, including modeling language, decision-making and knowledge management, life cycle integration, architecture modeling, and interoperability. From the perspective of these key topics, in this study, a systematic literature review was conducted to propose an ontology-based MBSE enabling framework. This framework consists of: (1) an ontology-based MBSE paradigm that reconfigures MBSE processes through enhanced semantic interoperability, optimized life cycle management, and dynamic decision support mechanisms; (2) a reference architecture that integrates MBSE models with ontology; and (3) key technologies that address semantic consistency verification and multidomain integration. Finally, the industrial applications are discussed as illustrative, preliminary, and qualitative evidence, suggesting the potential applicability of the proposed ontology-based MBSE framework in real-world engineering scenarios, particularly in the aviation industry. This work offers a structured overview of the current state, research hotspots, future trajectories, and enabling framework of ontology in MBSE. Future work can further investigate standardized evaluation metrics and benchmarking methods to quantitatively assess the effectiveness and performance of the OBMBSE framework in industrial applications.

Recent years, robot technology has made great progress benefiting from the breakthrough of sensor technology, new generation artificial intelligence, Information Communication Technology, etc., which results in more flexible, more dexterous and smarter robots for more complex application scenarios. Perception is the initialization of realizing specific tasks for robots, which should obtain enough information, understand the information, and make right decision. The emerging embodied intelligence provides a new paradigm for robots to deal with more dynamic and complex tasks, that is, embodied intelligent robots, where requires the the robots to be able to perceive more efficiently, more accurately, and smarter. As the behavioral pattern of human can be good guidance when handling different tasks, human-like perception gradually comes into view for embodied intelligent robots. To explore the development process and future trend of robot perception, a review from the perspective of human-like perception for embodied intelligent robots is carried out in this paper. The development trend towards to human-like perception for robot is analyzied, which the definition of human-like perception is provided. The technical path for realizing human-like perception is summarized, including multimodal sense, hierarchical understanding, and active decision and execution. The application potential and key challenges of human-like perception are discussed as well. It is a timely and significant review for the perception investigation of the emerging embodied intelligent robots, which could be an important reference for future works in this domain.

Hot stamping of 22MnB5 steel involves coupled heating, forming, and quenching stages, where variations in thermal exposure and cooling conditions can a ect austenite formation, phase evolution, and the nal mechanical response. In data-limited studies, process-window selection based only on point predictions may give an overcon dent assessment of feasibility. This work develops an uncertainty-aware process microstructure feasibility framework by linking a heating-stage surrogate, (HR,STT,ST) → γin, with a forming quenching surrogate, (γin,T0,∆T,CR,DA) → (m,b,f). Both links are calibrated using digitized experimental data from published hot-stamping studies and evaluated through cross-validation, with leave-onecooling-rate-out validation used to test the regime sensitivity of the forming quenching link,Phase-fraction predictions are constrained using clip-and-renormalize projection and an additive-log-ratio formulation to preserve valid martensite bainite ferrite compositions. Monte Carlo propagation is then used to convert surrogate outputs into probabilistic feasibility maps, con dence intervals, and failure-mode indicators. Because paired experimental measurements linking (m,b,f) to hardness and tensile strength are not available, the microstructure property stage is used only as an illustrative screening layer, not as a validated property model. The results indicate that, after su cient austenitization, cooling rate and initial blank temperature dominate the predicted candidate regions. The framework supports data-supported process-window screening, while practical property-window de nition still requires future paired microstructure property measurements under representative process conditions.

Under Industry 4.0, market demand is rapidly transitioning from a reliance on conventional mass customization to requiring mass personalized customization, calling for manufacturing systems to be equipped with reconfigurable machine tools (RMTs) that offer greater flexibility and efficiency and are driven by smart technologies. In general, the effect of demand fluctuations on the operation of smart manufacturing systems (SMSs) is subject to a prolonged propagation process, especially when the optimization of key production segments such as process planning, reconfiguration, and scheduling lacks organic integration, resulting in low efficiency and responsiveness. To tackle this issue, this article performs integrated optimization of process planning, reconfiguration, and scheduling for SMSs with RMTs, considering flexible process routes, adaptive RMT reconfiguration, and flexible scheduling. A deep reinforcement learning (DRL)-based approach is proposed to address this integrated optimization problem. A reward function based on penalty and prediction enhancement is devised to guide the learning progress, and a double deep Q-network (DQN)-based training framework is adopted to explore the best decision actions for each state. Finally, a real-world industry case study and extended numerical experiments are conducted to demonstrate the practical applicability and superiority of this integrated optimization approach. Moreover, to further explore the empowerment of industrial optimization through generative artificial intelligence (GAI), the potential of adopting large language models (LLMs) to address integrated optimization problems is discussed.

With the improvement of automobile intelligence, the research and development of the system is becoming increasingly complex and the development of the overall life cycle of the automobile system is facing increasing challenges. In this study, we propose a model-based systems engineering (MBSE) method through the development of multiarchitecture modelling language in order to support complex product development process modelling starting from the mission. Meanwhile, index verification and hybrid automaton simulations are employed to realise the static cost analysis of an architecture scheme and the rationality analysis of behaviour properties during the development process. Finally, a case of intelligent electric vehicle in a dual-vehicle system is established. In addition, the development process of mission-operation-feature/function-logic-physical (MOFLP) architecture is modelled and verified using multiarchitecture modelling language. Our results demonstrate that the multiarchitecture modelling language can support the multistage architecture description of the life cycle of complex products and perform static and dynamic analysis during the architecture description process.

Model-based systems engineering (MBSE) allows system models to formalize end-to-end systems engineering implementation while developing complex engineering system. The evolution of MBSE models, including changes and conflicts, provides important historical knowledge to support design decisions. Model versioning is an efficient approach to manage the evolution of MBSE models. However, the heterogeneous data structure and semantics used in MBSE practices hinder the tool interoperability that is required in model versioning, which also decreases the effectiveness and efficiency of system development. This paper proposes a tool-chain for model versioning of MBSE models based on a cognitive digital thread (CDT). In this tool-chain, the graph- object-point-property-relationship-role-extension (GOPPRR-E) modeling approach is adopted because it is compatible with heterogeneous modeling languages used in model versioning. To promote tool interoperability, this tool-chain adopts the Open Services for Lifecycle Collaboration to support conflict detection or resolution during model versioning. In particular, knowledge graphs are generated along with the model versioning workflow to develop a CDT, which provides the cognitive reasoning ability required for model versioning behaviors. A case study of landing gear system development is used to evaluate the feasibility of the proposed tool-chain through qualitative and quantitative analyses. The results demonstrate that the proposed tool-chain has better efficiency than traditional model versioning using Git tools.

Model reconstruction is a method used to drive the development of complex system development processes in model-based systems engineering. Currently, during the iterative design process of a system, there is a lack of an effective method to manage changes in development requirements, such as development cycle requirements and cost requirements, and to realize the reconstruction of the system development process model. To address these issues, this paper proposes a model reconstruction method to support the development process model. Firstly, the KARMA language, based on the GOPPRR-E metamodeling method, is utilized to uniformly formalize the process models constructed based on different modeling languages. Secondly, a model reconstruction framework is introduced. This framework takes a structured development requirements based natural language as input, employs natural language processing techniques to analyze the development requirements text, and extracts structural and optimization constraint information. Then, after structural reorganization and algorithm optimization, a development process model that meets the development requirements is obtained. Finally, as a case study, the development process of the aircraft onboard maintenance system is reconstructed. The results demonstrate that this method can significantly enhance the design efficiency of the development process.

In dynamic environments, such as box-pushing tasks, multi-agent systems (MAS) face significant challenges in coordinating agents within high-density settings while managing uncertainties arising from fluctuations in agent configurations and environmental dynamics. In this study, we explore the integration of surrogate response surface modeling (SRSM) with optimization algorithms, comparing Stochastic Gradient Descent (SGD) with a fixed learning rate, and adaptive learning rate (ALR) optimizers—including Adaptive Moment Estimation (ADAM) and Adaptive Approximate Direction Method Algorithm (AADMA)—to enhance MAS performance metrics, such as Agent Collision Rate (ACR), Agent Movement Frequency (AMF), and Task Completion Time (TCT). Through systematic experimentation across five scenarios, SRSM is employed to uncover key trends in MAS performance and identify configurations that improve scalability and adaptability. From the analysis of simulation data, it has been observed that SGD struggles significantly in dynamic environments, while ADAM demonstrates moderate improvements. However, AADMA consistently outperforms both by reducing loss, lowering collision rates, increasing movement efficiency, and achieving shorter task completion times. Performance comparison charts and loss function graphs emphasize AADMA’s superiority in addressing the complexities of real-time coordination and adaptability. Through this study, we highlight the critical role of combining SRSM with ARL to design an MAS that is capable of thriving in complex, dynamic, and high-density environments. By addressing key scalability and adaptability challenges, the proposed framework significantly advances MAS design, paving the way for improved multi-agent coordination in real-world applications.

Traceability plays an important role in model-based systems engineering (MBSE) practices, which helps engineers to ensure information consistency and correct constraint propagation throughout the system lifecycle. However, intelligently establishing trace links between heterogeneous models still faces several challenges, including those from tool interoperability and their inner scale complexity. To address these challenges, this paper employs a Cognitive Digital Thread (CDT) to develop a tool-chain to support intelligent trace links establishment between MBSE models. The CDT tool-chain first transform model information into unified services for tool interoperability using the Open Services for Lifecycle Collaboration (OSLC) standard. After that, a traceability engine are developed to intelligently capture and establish trace links based on model semantics in the proposed CDT tool-chain. A case study based on the design of a landing gear system demonstrates the feasibility and effectiveness of the proposed CDT tool-chain. The result of the case study indicates that the proposed CDT tool-chain improves the scalability and efficiency when establishing trace links between different MBSE models.

Recent years have witnessed unprecedented development in humanoid robotics, with dexterous hand grasping emerging as a focal research area across industrial and academic sectors. To track the state-of-the-art dexterous hand grasp, a review of dexterous hand grasp based on bibliometric analysis is executed. The related studies on dexterous hand grasp are collected from the Web of Science for analysis, where the publication details and cooperation situations from the perspectives of country, institute, etc. are discussed. The keywords cluster is adopted to find the hot research topic of dexterous hand grasp. The development trend of dexterous hand grasp is explored based on the top 25 keywords with the strongest citation bursts. The review findings indicate that precision control via multimodal fusion, autonomous task understanding and intelligent decision, and in-hand dexterous manipulation are top three hotspots in future.