Industrialized Construction (IC) is an approach to address ongoing challenges in the construction industry. Barriers affecting IC adoption include manufacturing inefficiencies caused by inaccurate production monitoring on shop floors. This paper presents the integrated use of machine learning and sensor data collected from IMUs and load cells installed on manufacturing platforms to monitor work progress. The research method includes the development of data collection, processing, and data analysis methods in a controlled laboratory, followed by validation on a real-world shop floor. Results present a 77% F1-score accurate monitoring through temporal windows and weighted deltas.
Human-Data Interaction (HDI) is a transformative research agenda that also represents a critical gap in current digital transformation efforts. Such transformation efforts have produced complex digital ecosystems that raise questions of, e.g., trust, agency, and ethics, but the predominantly technocentric approaches tend to prioritise system performance over human-centric aspects. The Built Environment (BE) is data-rich but insufficiently human-centred in how data are accessed, interpreted, governed, and acted upon. The HDI Committee, a permanent technical committee of the European Council on Computing in Construction (EC3 ), is dedicated to exploring the intricate relationship between humans and data in the BE. It investigates a wide range of topics, including user experience, change enablement, ethical data practices, and technological advancements that enhance our interaction with data. The Committee focuses on fostering dialogue and promoting leadership practices in HDI. It seeks to prioritise HDI-related research and educational initiatives, promoting practical applications across key areas.
Fireproofing spray on steel structures is essential for fire resistance, yet it is still manual, exposing workers to hazardous and dusty conditions. Robots have been developed to address these issues, but some systems depend on work area map, which limits real time adaptation. In addition, object level recognition cannot provide surface level work areas. Therefore, we propose an RGB D work area recognition model that classifies work area surfaces on a steel structure beam. On experiment, the model achieves a macro average F1 score of 0.8682. Finally, the proposed model provides surface level work area for fireproofing spray robot.
Current city-scale material stock assessment relies on archetype-based estimation with limited geometric grounding or on city-to-BIM conversion workflows that are costly and sensitive to representation detail. Both hinder auditable roof typology and quantity derivation from open 3D city models. This paper proposes an interpretable CityJSON pipeline that merges roof meshes into planar patches, extracts roof metrics (area, projection, slopes, span), performs rule-based roof classification (flat/shed/gable/hip/complex), and computes rule-based roof-layer and timber quantities. On 65 labelled buildings, LoD2 achieves 0.785 accuracy (macro-F1 0.705);BIM comparison on six buildings supports feasibility. The contribution is a reproducible, explainable roof-to-quantities method with quantified LoD sensitivity.
This study examines how virtual reality (VR) supported co-design methods can help recover children’s perspectives in shaping hospital environments. Building on NHS priorities for child-centred healthcare design, the project developed immersive digital replicas of key hospital spaces through 360° video, 3D models, and VR simulations. These were tested with children from local schools to explore their perceptions, reduce hospital related anxiety, and gather design preferences. Findings show that VR supported children engagement, familiarity with clinical spaces, and insights into sound, colour, layout, and wayfinding. The study shows the potential of VR as a participatory tool for inclusive paediatric healthcare design.
Construction planning relies on tacit knowledge that is rarely documented, leading to recurring knowledge loss and limited standardization. This paper proposes a neuro symbolic reinforcement learning framework for multi category construction relationship reasoning that formalizes construction execution relationships as explicit constraints. The framework combines synthetic Work Breakdown Structure generation, domain constrained relationship inference using large language model based reasoning, and policy learning via Group Relative Policy Optimization with symbolic rewards to produce explainable construction relationship graphs at inference time. Domain constraints serve a dual role by grounding reasoning during data generation and providing verifiable reward signals during optimization. Experimental results demonstrate stable learning behavior, with structural validity and reasoning quality exceeding 75 percent across categories, bounded policy divergence with KL approximately 0.22, and a truncation ratio reduced to 13.5 percent, supporting more efficient, consistent, and reusable automated construction planning across projects and organizational contexts.
Agentic AI describes a new generation of AI that automates tasks, pursues goals independently, makes decisions and proactively interacts with its environment. This study - based on a qualitative exploratory research design with expert interviews - aims at understanding how the use of such systems affects internal decision-making logic and organizational structures in companies. Results show that agentic AI changes decision-making processes. Decisions are being made by systems and humans, with humans mainly taking on supervisory or control functions. We provide a theoretical contribution to existing decision-making and organizational theories with practical recommendations for dealing with such autonomous AI systems.
Cyberattacks on critical infrastructure (CI) generate rapid and immediate public reactions that shape trust, risk perception, and crisis communication. Existing literature lacks an empirical evidence-based understanding of how decentralized online communities shape narratives following cyber incidents on CI. This study examines social media public response and discourse patterns to CI cyber disruptions through a reddit-based multi-dimensional sentiment analysis of Florida water-treatment-plant incident. Results from 300-relevant posts indicate that discourse is highly time-sensitive, 50%-participation and 90%-engagement occurred within six-hours, demonstrating rapid narrative formation. Findings highlight how online community signals can inform cybersecurity-communication strategies and support more transparent, trust-oriented CI protection efforts.
The transition from raw as-built data to semantic object representations is a bottleneck in construction digitalization, currently hampered by manual dependencies and the "black-box" nature of existing point cloud segmentation networks. This paper introduces an autonomous self-correcting 3D segmentation agent tailored for construction environments, utilizing 3D Gaussian Splatting (3DGS) as the underlying representation and Vision Language Models (VLMs) as the reasoning core. Unlike static segmentation methods, our agent adopts a dynamic approach by performing active scene roaming. Through a closed-loop feedback mechanism, the VLM analyzes visual data, detects objects, and rigorously assesses the plausibility of segmentation results. The agent iteratively corrects its own errors until a high-fidelity, logically consistent segmentation is achieved. This method demonstrates superior interpretability and accuracy, enabling downstream applications such as as-built verification, progress tracking, and quantity takeoff. By bridging generative rendering with semantic reasoning, our approach provides a robust framework for automated 3D scene understanding.
Fire safety compliance checks in existing buildings are often labor-intensive and prone to error, due to the lack of design data and the reliance on traditional/manual inspection tools. This work proposes an alternative to conventional approaches through a semi-automated Scan-to-BIM methodology to create a BIM model directly from point cloud data. This model is specifically designed to facilitate the regulatory checks required for fire safety, particularly at the level of compartmentation, and can be applied to evacuation routes in later stages. The methodology involves four main phases: i) point cloud acquisition and registration; ii) point cloud processing; iii) geometric features acquisition, and iv) creation of spatial entities compatible with the IFC format. The methodology demonstrates robustness when applied to benchmark datasets. The results confirm the efficiency and accuracy of the methodology in developing early design and BIM models suited to support fire safety engineering and verification of compliance with fire safety regulations.
