Spraying is a critical surface treatment process in intelligent manufacturing, and coating quality directly affects product performance. Therefore, efficient, accurate, and intelligent coating defect detection is an essential technique to ensure product reliability. The past decade has witnessed rapid progress in coating defect detection techniques. However, most existing studies have focused on specific methods or application scenarios, and there is a lack of systematic reviews that provide a comprehensive overview of this particular research area. To fill this research gap, this paper systematically reviews recent advances in coating defect detection, which covers methods from physical property-based non-destructive testing to deep learning-based approaches. Their fundamental principles, applicability in intelligent manufacturing, and current research progress are examined, and key challenges and potential solutions are discussed. Furthermore, integration of advanced intelligent manufacturing technologies into coating defect detection systems is analyzed to enhance system-level digitalization, automation, and efficiency. Finally, future development trends are explored and analyzed, including collaborative perception, cross-modal fusion, and autonomous decision-making. It is expected that this review will help to advance and accelerate theoretical research and engineering applications in coating defect detection by providing researchers with a comprehensive understanding.

Industry 5.0 advocates human-centric smart manufacturing (HSM), with growing attention to proactive human-machine collaboration (HRC). Meanwhile, the rapid development of Multimodal large language models (MLLMs) and embodied intelligence is driving an unprecedented evolution. This work aims to leverage these opportunities to enhance robots’ learning and cognitive capabilities, enabling seamless and natural interaction. However, current research often overlooks human–robot symbiosis and lacks attention to specialized models and practical applications. This review adheres to a human-centric vision, taking language as the pivot to connect humans with large models. To our best knowledge, this is the first attempt to integrate HRC, MLLMs and embodied intelligence into a holistic view. The review first introduces representative foundation models to provide a comprehensive summary of state-of-the-art methods in the ”Perception-Cognition-Actuation” loop. It then discusses pathways and platforms for efficient spatial skills learning, followed by an analysis of four key questions from the ”Why, How, What, Where” perspectives. Finally, it highlights future challenges and potential research directions. It is hoped that this work can help fill the research gap between HRC and MLLMs, offering a systematic pathway for developing human-centered collaborative systems and promoting further exploration and innovation in this exciting and crucial field. The resources are available at: https://github.com/WuDuidi/MLLM-HRC-Survey.

Worker resources are crucial in aircraft final assembly lines (AFAL), which are characterized by extensive manual assembly tasks. The features of AFAL, including resource constraints, makespan balancing, and flexibility in resource allocation, present greater challenges than conventional scheduling problems. This paper addresses the joint optimization problem of worker allocation under a resource dedication policy and scheduling of multi-mode tasks in the AFAL. Bi-objective with lexicographic order of minimizing the cycle time and total worker investment is considered, and an integer programming model is developed to formulate this problem. We propose a resource reallocation embedded genetic algorithm (RReGA) to solve this optimization challenge effectively. Initially, hybrid dispatch rules (HDRs) are employed to estimate the resource-makespan mapping of each workstation, yielding a high-quality initial resource allocation solution. Leveraging these mappings, a resource reallocation method, composed of a resource transfer strategy and a resource recovery strategy, is embedded in the evolutionary process of the genetic algorithm (GA) searching for scheduling solutions at the workstation. The resource transfer strategy is responsible for dynamic resource transfer across workstations, following a novel transfer principle to optimize the cycle time; while the resource recovery strategy aims to meet makespan constraints with the fewest workers to minimize cost. The efficacy and superior performance of the proposed algorithm are validated through comprehensive comparison and ablation experiments, as well as an unbalanced case study.

Robotic welding envisioned for the future of factories will promote high-demanding and customised tasks with overall higher productivity and quality. Within the context, robotic welding parameter prediction is essential for maintaining high standards of quality, efficiency, safety, and cost-effectiveness in smart manufacturing. However, data acquisition of welding process parameters is limited by process libraries and small sample sizes, given complex welding working environments, and it also requires extensive and costly experimentation. To address these issues, this study proposes a transfer learning and augmented data-driven approach for high-accuracy prediction of robotic welding parameters. Firstly, a data space transfer method is developed to construct a domain adaptation mapping matrix, focusing on small sample welding process parameters, and a data augmentation method is adopted to transfer welding process parameters with augmented sample data. Then, a DST-Multi-XGBoost model is developed to establish a mapping relationship between welding task features and welding process parameters. The constructed model can consider the relationship between the output, which reduces the complexity of the model and the number of parameters. Even with a small initial sample size, the model can use augmented data to understand complex coupling relationships and accurately predict welding process parameters. Finally, the effectiveness of the developed approach has been experimentally validated by a case study of robotic welding.

Industrial robots (IRs) serve as critical equipment in advanced manufacturing systems. Building high-fidelity digital twin models of IRs is essential for various applications like precision simulation, and intelligent operation and maintenance. Despite technological potentials of digital twins, existing modeling methods for industrial robot digital twins (IRDTs) predominantly focus on isolated domains. This fails to address inherent multi-domain complexities of IRs that arise from their integrated mechanical-electrical-control characteristic. To bridge this gap, first, this study proposes a multi-level multi-domain (MLMD) digital twin modeling framework and method. The framework systematically integrates physical space, digital space, and their bidirectional interactions, while explicitly defining hierarchical structures and cross-domain mechanisms. Subsequently, a four-step method is established, which encompasses component analysis, parameter extraction, MLMD IRDT modeling based on function blocks (FBs), and model validation. Then, implementation details are illustrated through an SD3/500 IR case study, where domain-specific modeling techniques and cross-domain integration mechanisms are systematically analyzed. Finally, effectiveness and feasibility of the proposed method is validated through experiments.