Recently, new technologies such as Building Information Modeling (BIM), laser scanning, and Lean construction approaches have significantly impacted the construction progress monitoring performance. However, mistakes during construction still occur due to drawing errors, poor communication, limited expertise, accidents, and other factors. To address this challenge, an improved Scan-vs-BIM methodology is proposed. Particularly, the Face-vs-Segment comparison incorporates semantic information, documents the magnitude and direction of deviations, and provides customized reporting. This approach reduces monitoring time, increase overall productivity, and improves decision-making processes. Thus with future research construction progress monitoring can be efficiently enhanced.
Automated generation of building information modeling objects from computer-aided design drawings require comprehensive information about material layers, including their functional roles—information typically absent from drawings. This study proposes a hybrid two-step methodology that combines a large language model and a machine learning algorithm to address it, focusing specifically on wall and slab objects. First, the method identifies the function of each material layer as a multi-class classification task, utilizing few-shot prompting with GPT-4o, achieving 99.8% macro F1-score. Second, it classifies whether material layer information pertains to walls or slabs using random forest model with FastText embedding, achieving 89.5% macro F1-score.
Semantic point cloud segmentation plays a role in creating geometric-semantic as-is models (GSMs), e.g., for planning and construction in existing contexts or creating digital twins. Deep learning architectures have proven useful in learning complex geometric patterns but focus on large context windows, neglecting fine-grained texture features by subsampling the input. In this contribution, we describe a feature extraction based on local point cloud neighborhoods and full-waveform terrestrial laser scanning data for material classification. Through sensitivity analysis, we show that this provides a useful basis for subsequent use in large-scale machine learning models.
Building circularity assessments aim to reduce waste and primary material demand. Indicators like the disassembly potential (DP) assess the detachability of elements and layers, but Building Information Modeling (BIM) lacks sufficiently detailed DP information in early design stages. Key information, such as connection types between elements and components, is missing in open BIM formats such as Industry Foundation Classes (IFC). To address this, we propose a new method using labeled property graphs to represent component-specific archetypes based on BIM. Tested in a case study, this approach enhances assessments of DP, providing designers with improved insights for early-stage circularity decisions.
Buildings own diverse datasets, including geometric, product, logistic, real-time monitoring, regulatory, and occupant feedback data. However, challenges such as data scarcity, insufficient labelling, and the complexity of multimodal data limit conventional AI’s ability to provide accurate, scalable and content-aware insights, often confining its application to specific buildings and time. General-Purpose Artificial Intelligence (GPAI) offers the transformative potential to maximise the value of data. Early research explores adaptive AI, meta-knowledge transfer, synthetic data, and foundation models to support generalisation across tasks. This paper examines how these developments position GPAI as a step toward general-purpose intelligence in buildings.
Traditional education models delivered remotely struggle to sustain student engagement. However, novel learning approaches are rarely adopted at scale due to institutional resistance, lack of structured implementation frameworks, and misalignment with existing curricula. This paper aims to examine the impact of incentivised learning through the integration of digital badges as motivational tools within a Virtual Learning Environment (VLE), to encourage self-paced, autonomous learning. Using a mixed-methods approach, quantitative data were collected through student performance tracking, and model accuracy measurements, alongside qualitative feedback from surveys. Findings indicate that achievement-driven incentives enhance student and educator motivation, engagement, and accelerate skill acquisition.
Ultra-wideband (UWB) technology, offering high precision and low latency, emerges as a solution for accurate worker tracking to improve construction safety. However, non-line of sight (NLOS) conditions can impact its reliability. This study evaluates UWB tag placements on different body parts in a controlled walking experiment. Data from six participants were analyzed using t-Tests for accuracy and F-tests for consistency across different axes. Results show that head-mounted tags provide the highest accuracy, while tags attached around shoulders, waist, and knees deviate due to body shadowing. Findings emphasize the importance of optimal sensor placement for real-time worker tracking in construction.
This study investigates Digital Transformation in the AECO industry through the lens of ‘Ikigai’, which is a concept signifying the intersection of passion, mission, vocation and profession. The research adopts an abductive, qualitative approach, conducting 4 semi-structured interviews with industry experts. Findings reveal that DT strategies should be purposeful to deliver productivity and efficiency, better outcomes, sustainability, specialism, and proficiency in the delivery of products and services in the AECO industry. The study proposes a preliminary conceptual framework that integrates the ‘Ikigai’ philosophy into DT strategies, advocating a platform perspective to unify digitally-enabled deliverables with sustainable and purpose-driven practices.
This paper outlines the development of a Existing Building Digital Twin system for social housing in Scotland. The project focuses on enhancing retrofit planning and execution. Research methodology involved stakeholder workshops, surveys, and case-based reasoning methods. Key findings highlight the need for a decentralised, secure digital system integrating diverse data sources. The paper proposes ten components addressing various aspects of DT development, aiming to create a robust, user-friendly tool for decision-making in retrofit projects, supporting the country's transition to net-zero carbon emissions. The system is expected to offer benefits including improved energy efficiency, cost savings, and enhanced living conditions.
