This survey provides a structured analysis of intelligent assembly support systems, examining how these systems perceive tasks, reason about assembly processes, generate support actions, and interact with human operators in complex industrial environments. A systematic literature review of 184 publications from 2015 to 2026 was conducted, combining thematic analysis with bibliometric and temporal analyses to identify research trends, technological developments, and persistent challenges across perception, reasoning, interaction, and deployment. The analysis is structured around three perspectives: perception and task understanding under uncertain industrial conditions, adaptive support for heterogeneous workers, and emerging technologies with implications for industrial deployment. The review reveals that current systems remain predominantly focused on vision-based perception, action recognition, and predefined support strategies, with limited consideration of perception uncertainty arising from visual ambiguity and environmental changes, variations in human assembly behaviour, and diverse worker capabilities and support needs. By synthesising fragmented research streams across technical, human, and deployment perspectives, this survey identifies a fundamental transition from recognition-oriented support tools towards uncertainty-aware decision-support systems that enable adaptive human–AI collaboration. Key challenges include robust perception under environmental and visual variability, adaptive support for heterogeneous workers, trustworthy integration of emerging technologies, such as large language models and human-centric digital twins, and practical requirements for industrial deployment. This review establishes a broader perspective for evaluating intelligent assembly support systems, proposing that future progress should be assessed not only through technical performance but also through adaptability, trust calibration, human autonomy, and long-term operational viability. The findings provide a roadmap for Industry 5.0 assembly environments in which intelligent systems augment rather than replace human expertise.
The behaviour of humans when collaborating with other actors such as other humans or robots has been studied for decades and remains critical for unlocking mixed human-robot team performance. To achieve better, more effective human-robot collaboration (HRC) in industrial settings, a robot needs to have robust models to predict human behaviour when working towards a shared goal. Inferring and modelling human intentions is expected to play a pivotal role in accurately predicting human behaviour as suggested in this paper. This paper explores the potentials, challenges, and opportunities in human behaviour and intention prediction for industrial HRC. It first examines human factors in socio-technical systems, highlighting the challenges of human-human collaboration that inform the design of human-robot teams. Building on this foundation, a human behaviour and intention prediction framework is introduced that enables seamless HRC by integrating behavioural modelling techniques. The paper further reviews perception systems used to interpret the physical world and decode physiological and affective states, along with reasoning and planning strategies based on probabilistic, rule-based, machine learning, hybrid, and language model approaches. Verification and validation methodologies are discussed to ensure reliability and safety, followed by an analysis of the ethical, legal, and societal implications of predictive modelling in industrial HRC. Finally, the paper concludes with a discussion of open challenges and future research directions, emphasising advancements in adaptive modelling, perception-driven reasoning, and ethical frameworks to shape the next generation of collaborative robots that are robust, human-centric, and ethically responsible. Data Availability No new data were generated or analysed in this study.
Manufacturing companies are facing increasing pressure on streamlining assembly processes while maintaining sustainable product lifecycles. Well-established tools such as Design for Assembly (DFA) and Design for Disassembly (DfD) provide guidance for simplifying assembly and product lifecycles, but their unified strategic use evaluated empirically and holistically across assembly efficiency, feasibility of automation and disassembly performance in a single study remains insufficiently explored. This study investigates how DFA principles can be applied in early-stage product development processes, for simplifying assembly by reducing complexity, improving the feasibility of automation and influencing disassembly performance. This study adheres to Design Science Research Methodology, using drone assembly as a case, targeting an initial drone version resembling many commercially available products contrasting a redesigned drone based on DFA principles. Results show that applying DFA principles results in considerable reduction of drone parts and components contributing to 49% reduction in assembly time, and 53% reduction in disassembly time. It also improves the feasibility of automation through standardised part geometry and modular design. The study demonstrates DFA's capability as a strategic enabler of efficient assembly processes but shows that it must be carefully balanced with DfD principles to enable sustainable and circular lifecycles. Findings provide valuable guidance for researchers, engineers, and designers planning to incorporate circularity principles, or enhance efficiency of assembly processes in product and production development.
