Rapid and accurate automatic fall detection is essential for improving worker safety and reducing the severity of fall-related incidents on construction sites. To address the challenges of real-time detection in complex and obstructed construction environments, this paper develops a specialized dataset for fall scenarios and introduces a skeleton-based AI model called YOSAP-LSTM. This model integrates

With the rapid development of artificial intelligence, intelligent algorithms for parameter nonlinear mapping provide a new design approach to address the long-term reliance on empirical design in tunnel engineering. This paper proposes an intelligent model for predicting support structure parameters based on tunnel background information. After comparing the characteristics of machine learning and

Recent advancements in AI-based Digital Twins (DTs) have substantially influenced bridge monitoring and maintenance, especially through Deep Learning (DL) for sensor-based damage detection. However, the effectiveness of DL models is constrained by the extensive training data they require, which is often costly and time-consuming to collect in bridge infrastructure contexts. To address this data scarcity

Training workers in safety protocols is crucial for mitigating job site hazards, yet traditional methods often fall short. This paper explores integrating virtual reality (VR) and large language models (LLMs) into iSafeTrainer, an AI-powered safety training system. The system allows trainees to engage with trade-specific content tailored to their expertise level in a third-person perspective in a non-immersive

3D multibeam sonar is a feasible solution for detecting bridge scour. However, the reliance on technicians for the identification of the morphological characteristics of local scour pits is time-consuming and subjective, and the absence of surface data hinders scour morphology analysis. Hence, an algorithm is proposed for the unsupervised identification and precise reconstruction of bridge scour morphology

The construction industry has experienced significant transformation with the adoption of Building Information Modeling (BIM) technology, which provides a virtual representation of both the physical and functional components of construction projects. In this context, project scheduling becomes a vital tool for site managers to assess construction progress. This paper introduces an augmented reality

Building performance modeling and analysis using Building Information Modeling (BIM) platforms remains fragmented, requiring various software applications to address different disciplines. Challenges in data extraction, transfer, and integration arise due to inconsistencies in vendor-specific data schemas and limited interoperability. Moreover, OpenBIM data schemas lack comprehensive object definitions

Motor graders are heavy construction equipment that specialize in leveling ground surfaces. These machines adjust the blade beneath the vehicle body to perform various tasks. With the growing interest in the automation of motor grader operations, the complexity of the earthmoving mechanism and multitasking while driving degrade the efficiency and performance of these operations. Controlling the blade

Raw point clouds collected from real-world scenes are sparse, incomplete and noisy, posing significant challenges for their integration into automation workflows in construction. Thus, completing plausible and fine-grained point clouds is a critical prerequisite for downstream applications. Current methods primarily focus on learning patch-level features and modeling their relationships for inferring

Fragmentation and poor collaboration in contract-heavy industries hinder innovation. While smart contracts offer promising automation for digital documents, the transformation process presents significant challenges. Current approaches are promising but are often constrained by technical limitations, domain-specific requirements, and limited flexibility, restricting widespread adoption. This paper

The inspection of large-scale tunnel networks is essential to identify any long-term deterioration mechanisms. Traditional manual inspection is not automatic and adaptive in geometrically complex tunnel configurations. This paper presents an automatic unmanned aerial vehicle (UAV)-based tunnel assessment application in a critical cross-passage twin tunnel section of Dublin Port Tunnel. It adopts a

Automating 3D reconstruction of indoor environments is essential for scene understanding in Building Information Modeling (BIM). This paper addresses the challenge of integrating geometric and material attributes in scan-to-BIM processes. A deep learning-based framework is developed to automatically extract and integrate geometric and material attributes from point clouds, incorporating an enhanced

The flows of materials arriving at, and waste moving out of a site, require detailed planning for smooth logistic processes. This is underscored by the temporal and spatial limitations around sites, characterized as a pure bottleneck. This paper proposes a conceptual data framework for developing digital twins for logistics capacity planning in construction to align the capacities for (i) incoming

“VR + BIM” technologies address the growing complexity of construction evaluation systems driven by digital transformation, as traditional methods struggle to meet diverse stakeholder demands. This paper develops an automated evaluation framework integrating Building Information Modeling (BIM) and Virtual Reality (VR) to enhance decision-making processes and enable immersive multi-stakeholder collaboration

Bridge condition assessment is often compromised by errors in inspection data, limiting reliable maintenance and management decisions. This paper investigates how to enhance inspection data quality by automatically identifying and correcting the inaccurate assessment of structural conditions. A model that integrates textual and quantitative features is proposed to identify defect and condition ratings

An effective and efficient review of construction contracts is essential for minimizing construction projects losses, but current methods are time-consuming and error-prone. Studies using methods based on Natural Language Processing (NLP) exist, but their scope is often limited to text classification or segmented label prediction. This paper investigates whether integrating Large Language Models (LLMs)

Steel reinforcement design is essential for the structural integrity and durability of reinforced concrete (RC) structures. However, rebar clashes frequently occur due to conventional design processes lacking precise bar positioning, leading to time-consuming and error-prone onsite modifications. Existing 3D analysis tools for clash detection are unsuitable for rebar design, which must comply with

During the inspection and maintenance of underground pipe systems, technicians often spend considerable time searching for subtle defects in inspection videos captured under varying pipe conditions using Closed-Circuit Television (CCTV). The lack of feature extractors tailored for pipe images, combined with the complexity of pipe CCTV videos, poses substantial challenges to the performance and applicability

Existing technologies struggle to accurately identify interactions between workers and equipment, as well as the deep semantics of complex construction scenes. To address these limitations, this paper proposes an automated construction site safety management system designed to enhance scene understanding and identify safety hazards while focusing on hazard-area and personal protective equipment (PPE)

Smart infrastructure management requires continuous monitoring and acquisition of heterogeneous data from diverse sources and their fusion for effective predictive maintenance. However, existing methods lack multi-level fusion, are limited to single-class outputs, and fail to incorporate maintenance actions alongside bridge condition ratings. Hence, this research explores a deep learning-driven multi-level

Human activity recognition (HAR) in construction has gained attention for its potential to improve safety and productivity. While HAR research has shifted toward vision-based approaches, many studies typically use data from a specific angle, limiting understanding of how camera angles affect accuracy. This paper addresses this gap by using AlphaPose and Spatial-Temporal Graph Convolutional Network

Bolt loosening in steel bridges poses critical safety risks. This paper proposes a multi-view spatiotemporal framework to assess bolt loosening. First, YOLO detects gusset plates and bolts, with spatiotemporal correlation model extracting region-specific bolt video clips. An enhanced UNet architecture quantifies bolt shadow areas as loosening indicators. To address single-frame feature limitations

Artificial intelligence (AI) has significantly advanced infrastructure monitoring, particularly through machine learning and deep learning techniques. In bridge management, combining AI with Building Information Modeling (BIM) and unmanned aerial vehicles (UAVs) enhances accuracy, efficiency, and safety. This paper reviews AI, UAV, and BIM applications, focusing on technology integration and algorithm

The selection of data modelling methods in the data-driven performance prediction of tunnel boring machines is a challenge since each method has its own advantages and disadvantages compared with each other. Extreme learning machine (ELM) exhibits the benefits of fast learning speed, better scalability, and generalization performance, and is easy to convert between neural networks-based and kernel

Watermain failures universally challenge urban infrastructure, causing water loss, service disruptions, and maintenance challenges. There is a lack of research attention towards semi-tropical regions, which have unique environmental conditions affecting the deterioration of watermains differently. This paper develops a big data-driven prediction model for watermain failure types (no leak, leak, no

Leading edge erosion of wind turbine blades reduces energy production and blade lifetime, a growing issue with larger blades. Effective monitoring is crucial to tracking erosion and controlling maintenance costs. This paper presents an image-based 3D reconstruction method for the leading edges, targeting challenges like textureless surfaces, background motion, and limited image overlap that cause existing

Driven by Industry 4.0, the construction of reservoir dams is entering a new phase characterized by both opportunities and challenges. The deep integration of technologies such as artificial intelligence, big data, and digital twins aims to enhance the safety, quality, efficiency, and sustainability of dam projects. This paper utilizes VOSviewer analysis software to conduct a bibliometric analysis

Traditional pavement repair techniques are time-consuming, labour-intensive, prone to errors, and expose manpower to high-risk road traffic conditions. This paper proposes a data-driven solution for planning and automating the repair process for road potholes using a fleet of unmanned ground vehicles (UGVs). The project encompasses data mining, developing software tailored for fleet management, and

This paper presents a systematic review of generative artificial intelligence (AI) use in architectural design from 2014 to 2024, focusing on 1) AI models and theory-application gaps, 2) design phases, tasks, and objectives, 3) data types and contents, and 4) evaluation methods. Based on 161 journal papers selected using preferred reporting items for systematic reviews and meta-analysis (PRISMA), the

The Architecture, Engineering, Construction, and Operations (AECO) sector gains significant advantages through generating and effectively managing BIM data. This increased available data can be fundamental in deriving innovative advances by processing them through artificial intelligence (AI) models. In this context, this paper investigates how BIM and AI capabilities can benefit the development of

Accurate detection and comprehensive assessment of corrosion states are essential for bridge safety and durability. Deep learning-based semantic segmentation methods show significant potential for corrosion detection. However, supervised methods confront substantial challenges in labor-intensive annotation and limited datasets. To address these challenges, a semi-supervised method for corrosion state

Deep learning methods for sewer defect detection face challenges due to their reliance on time-consuming bounding box annotations and lack of model interpretability. This paper proposed a framework leveraging weakly supervised object localization (WSOL) that requires only image-level annotations. Analysis showed that effective performance could be achieved with minimal training data (100 images per

Aging bridges within regional transportation networks often suffer from structural deficiencies, requiring effective maintenance strategies under budget constraints. This paper aims to determine optimal maintenance strategies at the system level using a semi-Markov process-based dynamic optimization approach. The proposed method models bridge condition evolution with a semi-Markov matrix and integrates

Effectively managing inspection and maintenance data in facility management remains challenging due to the lack of structured and interoperable data representation. This paper explores how the IFC schema can be leveraged to standardize inspection, damage, maintenance, and maintenance cost data representation. To this end, an IFC-based framework was developed to ensure semantic consistency and interoperability

A direct relationship exists between asphalt concrete (AC) density and pavement performance. In the United States, quality thresholds have been established for AC density by Departments of Transportation (DOTs). During flexible pavement construction, remedial actions become limited as AC cools. Therefore, it is crucial to monitor AC to achieve the desired density during compaction. Ground-penetrating

The conversion of image coordinates to physical coordinates, usually through camera calibration, is necessary to obtain the accurate displacement when using the vision-based measurement system. This paper proposes a vision-based damage identification for beam-type structures using area scanning without calibration. First, the modal parameter identification using the uncalibrated pixel displacement

Efficient analysis of road pavement distresses is crucial for infrastructure management and safety. Traditional methods are labor-intensive, and recent deep-learning approaches face challenges such as overlapping bounding boxes and poor pixel localization. This paper presents PaveGAN, a real-time method to identify and annotate pavement distresses using generative adversarial networks. While trained

In the Architecture, Engineering, and Construction (AEC) industry, design processes remain fragmented across architectural, structural, and mechanical domains, limiting integration and optimization opportunities throughout building lifecycles. This paper investigates how artificial intelligence can be leveraged to create a comprehensive framework for parametric generative design in industrialized construction

The absence of an efficient maintenance method has incurred substantial additional costs, emerging as the primary impediment to the advancement of marine reinforced concrete (RC) structures. This paper proposes a parametric modeling and evolutionary optimization method to improve the cost-effectiveness ratio of structural maintenance. The deterioration risk distribution of the entire structural system

Named Entity Recognition (NER) is a fundamental task for automatically processing and reusing documents. In traditional methods, machine learning has been used relying on costly high-quality datasets. This paper proposed an NER method based on fine-tuning Large Language Models (LLMs) with low-quality datasets for construction documents. Firstly, low-quality datasets were semi-automatically generated

In recent years, the advancement of digital technologies such as building information modeling (BIM), internet of things (IoT), and geographic information system (GIS) has had many impacts on the construction industry. However, limited research has been conducted on the integration of BIM, IoT, and GIS technologies, especially in construction resource monitoring. Therefore, this paper presents a state-of-the-art

Current expert-based approaches to determining the weights of different safety management elements during contractor safety performance are time-consuming and potentially biased.Hence, this paper evaluates analytics-based approaches, i.e., supervised learning, cluster-then-predict and two-level variable weighting K-Means (TWKM) (an extension of the traditional K-Means clustering algorithm), against

Bridges are susceptible to severe damage from natural disasters, heavy traffic loads, and material degradation, necessitating timely and accurate information for effective emergency response. Current bridge disaster management systems often fail to meet real-time requirements due to interoperability challenges and fragmented functionalities across different phases. This paper systematically reviews

The 3D wireframe model provides concise structural information for building reconstruction. Traditional geometry-based methods are prone to noise or missing data in 3D data. To address these issues, this paper introduces Edge-NeRF, a 3D wireframe reconstruction pipeline using neural implicit fields. By leveraging 2D multi-view images and their edge maps as supervision, it enables self-supervised extraction

Reasonable shape and cross-section design of free-form Single-Layer Reticulated Shells (SLRSs) are crucial for their superior static performance and material efficiency. However, traditional metaheuristics face high computational costs and are prone to converging to local optima when optimizing these factors simultaneously, often leading to necessity of carrying out decoupled design processes. This

Large-scale earthwork transportation encounters queuing congestion and dynamic uncertainties, while existing methods ignore complex traffic behaviors and exhibit limited responsiveness and generalization. This paper proposes a multi-task Deep Reinforcement Learning (DRL) framework for the dynamic scheduling of large fleets across supply sites and traffic networks. In the framework, multiple agents

Texturing Building Information Model (BIM) with up-to-date Unmanned Aerial Vehicle (UAV) images has brought substantial benefits to building façade inspection. However, current image-to-BIM registration methods are sensitive to UAV positioning accuracy and façade features. Additionally, perspective and geometry distortions on UAV images hinder the texturing of curved façades. To address these issues

The acquisition of pavement distress images using UAVs presents unique challenges compared to ground-based methods due to differences in camera configurations, flight parameters, and lighting conditions. These factors introduce domain shifts that undermine the generalizability of segmentation models. To address these limitations, an interactive segmentation model, CDCR-ISeg, is proposed to bridge the

Accurate monitoring and visualization of urban road substructure and targets are impeded by challenges in inverting Ground Penetrating Radar (GPR) data, especially under multiple inversion objectives and complex road conditions. To address this challenge, a deep learning-based multi-scale inversion approach, termed MSInv-GPR, is proposed, which builds on the Pix2pix Generative Adversarial Network (Pix2pixGAN)

Generative design has its roots in the 1990s and has become an intense research topic for bringing the power of artificial intelligence to various aspects of architecture practices. The recent advancements in artificial intelligence have made a methodological shift in innovative approaches to generative design, fueled by the proliferation of big data. This paper provides a comprehensive review of emerging

Despite advancements in computer vision techniques like object detection and segmentation, a significant gap remains in leveraging these technologies for hazard recognition through natural language processing. To address this gap, this paper proposes VQA-RESCon, an approach that combines Visual Question Answering (VQA) and Referring Expression Segmentation (RES) to enhance construction safety analysis

Worker-robot collaboration (WRC) has emerged as a transformative approach to augmenting the productivity of the construction industry. However, the development of a safety monitoring method or system for stopping robot operations in emergency is imperative, especially for seamless WRC on site. This paper presents a digital twin-enabled safety monitoring system for seamless WRC on site, characterized

The intelligence of electric shovels plays a critical role in improving excavation efficiency and safety. A key challenge in intelligent excavation is generating an optimal excavation trajectory while considering material uncertainty. Therefore, an Unmanned mining Electric Shovel Trajectory Planning method based on the Point-by-point Incremental B-spline Curve under Uncertainty (UESTP-PIBCU) is proposed

Pavement preventive maintenance is one of the most fundamental use cases when deploying digital twins (DTs) for highway infrastructure management. To achieve this, it is essential to accurately predict the pavement conditions in future years. This paper developed a Spatial-Temporal Graph Attention network (STGAT) that can effectively capitalize on both spatial and temporal dependencies while addressing

Machine vision sensors face challenges in automating the monitoring of internal structural damage and deformation, with limited lifespan and resolution accuracy. This paper develops a high-durable machine vision strain sensor, MISS-Silica. The sensor's durability is enhanced through materials, processes, and algorithms, ensuring its lifespan aligns with that of the structure. It combines an endoscope

Worker safety has become an increasing concern in human–robot collaboration (HRC) due to potential hazards and risks introduced by robots. Deep Reinforcement Learning (DRL) has demonstrated to be efficient in training robots to acquire complex construction skills. However, neural network policies for collision avoidance lack theoretical safety guarantees and face challenges with out-of-distribution

Layer pressing in 3D concrete printing (3DCP) allows to continuously modify the thickness of printed laces by changing adequately the robot speed. However, most applications consider a constant thickness throughout the printing and do not leverage all the possibilities from robotic technologies. The aim of this paper is to demonstrate the potential offered by thickness variation to achieve higher structural

Construction process monitoring traditionally relies on manual inspections and document cross-referencing, leading to inefficiencies in project management. Despite advances enabling computer vision-based monitoring and automated document analysis, integrating these technologies remains challenging, particularly in connecting field data with work documentation. This paper proposes an automated monitoring

Modular integrated construction (MiC) offers improved sustainability and automation. Nevertheless, its performance is impeded by extensive logistics operations, including multimode transportation, recurring loading-unloading, stacking, and assembly. Such rigorous operations may cause inadvertent underlying damage to module structure, leading to supply chain disruptions, safety hazards and structural