Integration of AI-Based Ergonomics in Early-Stage Building Design: A Frame Work for Enhancing Occupational Health in Construction Projects

Integration of AI-Based Ergonomics in Early-Stage Building Design: A Frame Work for Enhancing Occupational Health in Construction Projects

ORCID

(Arlovia Acevedo Noboa1)

Author: Arlovia Acevedo Noboa

Civil Engineer Occupational Health and Safety Specialist Troya Railway Constructions

EURING (Engineers Europe)

ASSP (American Society of Safety Professionals) WORK FORCE SAFETY HUB

Madrid, Spain

Abstract

This article presents a novel framework for incorporating artificial intelligence (AI) algorithms, specifically machine learning and computer vision, during the early design stages of architectural and engineering processes, aimed at predicting ergonomic risks in construction projects. [1]

[7] [13].

.

Unlike existing approaches that consider ergonomics only during the construction or operational phases, this proposal moves prevention forward to the design stage, when spatial and functional decisions already have a direct impact on occupational health [4] [8] [9]

The system merges digital modeling with data mining tools applied to posture analysis, physical loads, and areas of potential ergonomic risk, integrating these into Building Information Modeling (BIM) platforms commonly used in construction [1] [7] [13]. The strategy also aims to prevent work-related accidents, injuries, and fatalities, thereby improving productivity and reducing costs in civil construction [11] [12].

It is flexible in its alignment with various international standards and scalable at a global level [6] [10]. This research contributes to a broader and more proactive vision of engineering design, in line with the principles of human and technological sustainability promoted by the Journal of Engineering, Design and Technology.

Keywords

Artificial Intelligence, Ergonomics, Early-Stage Design, Occu- pational Health, Construction Safety, Predictive Modeling

The construction sector remains one of the industries with the highest risks to occupational health [12] [11]. In particular, work-related musculoskeletal disorders (MSDs) are a leading cause of absenteeism, reduced productivity, and increased medical costs. Despite advancements in technology and safety regulations, ergonomic risks are still addressed reactively typically late in the construction process or only once issues have already arisen on site.

Meanwhile, artificial intelligence (AI) is being increasingly adopted across engineering and design disciplines. Software applications in the industry commonly support structural analysis, energy optimization, and project management. However, their use as tools to predict ergonomic risks during the early design phase where spatial, geometric, and func- tional decisions can still be modified remains in its infancy or is nearly nonexistent [7].

This paper introduces a novel methodology that integrates AI technologies with ergonomic principles to identify and pre- vent postural and biomechanical risks at the earliest stages of architectural and engineering design. This not only facilitates safer working environments but also supports productivity and long-term sustainability in the construction sector.

The primary objective is to establish a methodological framework for incorporating AI algorithms (e.g., machine learning and computer vision) within digital modeling envi- ronments (e.g., BIM) to enable early prediction of potentially hazardous ergonomic conditions before construction begins. Such approaches promote a preventive, scalable, and globally adaptable solution in compliance with standards such as ISO 11226 [6], OSHA [10], and national occupational health guide-lines.

1. INTRODUCTION

   Figure 1. Prevalence of Musculoskeletal Disorders (MSDs) in the Construc- tion Industry. Rate of MSDs in the construction sector compared to other industries. Source: Project Safety Journal.

2. THEORETICAL FRAMEWORK

   1. Ergonomics in Construction: Ongoing Challenges Ergonomics in the construction industry focuses on adapting working conditions to the capabilities and limitations of the

      human body. Numerous studies have reported a high preva- lence of work-related musculoskeletal disorders (MSDs) among construction workers, particularly during tasks in- volving repetitive loads, awkward postures, or manual mate- rial handling [11] [12].

      Evaluation methods such as REBA (Rapid Entire Body As- sessment), RULA (Rapid Upper Limb Assessment), and the NIOSH Lifting Equation are widely used to assess ergonomic risks [4] [8] [9]. However, these tools are predominantly applied manually and reactively often after workers are already on-site and operating under fixed conditions.

      Figure 2. Distribution of risk factors for work-related musculoskel- etal disorders (WMSDs) causing work absence in the construction sector (2015). Overexertion during lifting/lowering.

   2. Artificial Intelligence in Architectural and Engineering Design

      In the fields of architecture and engineering, artificial intel- ligence (AI) has gained attention for its applications in au- tomatic floor plan generation, energy-conscious design, structural analysis, and sensor data interpretation [2] [3] [13].

      Machine learning techniques, neural networks, and evolu- tionary algorithms have demonstrated their capacity to rec- ognize complex patterns and propose adaptive design solu- tions. Nevertheless, the use of AI to predict ergonomic risk factors before a physical workspace has been constructed remains an emerging and largely unexplored area. While some preliminary studies have explored computer vision or biomechanical modeling to assess posture risks, these ap- proaches remain fragmented and rarely integrated into main- stream design workflows [1] [7] [13].

      Figure 3. Overview of work-related musculoskeletal disorders (WMSDs), their main symptoms, affected body parts, contributing factors, and prevention strategies. Source: European Agency for Safety and Health at Work (EU-OSHA), 2020.

   3. Digital Modeling and Human-Centered Design

      With the rise of Building Information Modeling (BIM) and other 3D modeling systems, new opportunities have emerged to integrate ergonomics and occupational health considera- tions into the digital design workflow [1] [7] [13].

      Human-centered design places emphasis not only on how a building looks or performs but also on how it affects the health, safety, and efficiency of the people who build or use it. This multidisciplinary approach enables the correlation of geometric, spatial, and functional parameters with biometric and biomechanical data, establishing a more holistic per- spective on risk prevention [5] [11].

   4. Limitations of Current Approaches and Integration Op- portunities

      Although some initiatives exist to digitize ergonomic as- sessments, there is a lack of systematic strategies for applying AI algorithms to forecast risks during the design phase. This gap represents a major opportunity to develop predictive tools capable of simulating tasks, analyzing posture, and utilizing anthropometric data to recommend ergonomic improvements and reduce risk exposure [1] [4] [9].

      The technical and financial feasibility of such an approach is supported by current digital modeling technologies, and its implementation could significantly enhance compliance with international occupational safety regulations [6] [10].

3. METHODOLOGY

   Building uon the theoretical insights presented in the pre- vious section, this paper proposes a four-phase methodology for integrating artificial intelligence (AI) algorithms into ergonomic evaluation during the architectural and engineering design stages. The focus is on digital modeling environments particularly BIM where construction tasks, postures, and workflows can be simulated. In this way, ergonomic risks can be detected and mitigated before implementation on the construction site [1] [7] [13].

   The project is initially developed using a three-dimensional model with BIM tools such as Autodesk Revit or ArchiCAD [13]. In addition to architectural and structural components, the model includes access zones, working heights, material han- dling areas, and construction sequences. This framework supports the creation of spatially realistic contexts for simu- lating actual construction scenarios [1] [7].

   1. Phase 1: Digital Modeling of the Work Environment

      The project is initially developed using a three-dimensional model with BIM tools such as Autodesk Revit or ArchiCAD. In addition to architectural and structural components, the model includes access zones, working heights, material han- dling areas, and construction sequences. This framework supports the creation of spatially realistic contexts for simu- lating actual construction scenarios [1] [3].

   2. Phase 2: Capture and Generation of Ergonomic Data

      AI models are fed with data from multiple sources, includ- ing:

      • International anthropometric databases (e.g., ISO – 7250 or regional datasets) [5] [6].

      • Posture simulations using tools like Jack (Siemens) or the AnyBody Modeling System.

      • Virtual environments where human avatars perform simu- lated construction tasks.

        The study focuses on tasks such as lifting, collecting, carry- ing, and working at heights [4] [8] [9].

   3. Phase 3: Development and Implementation of AI Algo- rithms

      Supervised machine learning classifiers (e.g., random forests or neural networks) are trained with datasets labeled accord- ing to ergonomic risk levels based on REBA, RULA, or NIOSH criteria. These datasets include synthetic data gener- ated through digital human modeling tools (e.g., Jack, Any- Body), complemented by publicly available anthropometric databases such as ISO 7250 and regional sources [5] [6]. These algorithms are trained to predict risk associated with indi- vidual design configurations such as element heights, material locations, or circulation paths.

      In addition, computer vision techniques are integrated to assess joint angles, forced postures, and estimated exposure times. The outputs are visualized as heatmaps overlaid on the BIM model, highlighting ergonomic risk zones [1] [13].

      Figure 4. Example of integrated ergonomic risk evaluation using biome-chanical simulation. Joint contact forces, NIOSH lifting index, RULA, and REBA analyses are shown for a lifting task, supporting real-time risk pre-diction in digital human modeling environments.

   4. Phase 4: Validation and Feedback into the Design

      The algorithm outputs provide actionable feedback for the design team, including automatic recommendations such as modifying heights, adjusting assembly sequences, or rede- signing access routes [1] [13]. Initial validation involves com- paring the AI-predicted results with expert human ergonomic assessments to evaluate accuracy, sensitivity, and

      practical relevance [4] [8].

      This methodology aims to promote preventive deci- sion-making during design, treating occupational health as a design variable as critical as functionality or energy efficiency right from the projects inception [7].

4. APPLICATION IN THE CONSTRUCTION SECTOR

   The proposed method represents not only a technical innova- tion but also a practical response to the urgent needs of the construction industry, which demands early identification of ergonomic risks and proactive mitigation to ensure safe and productive working conditions [1] [11] [12].

   Below are three key domains where AI-assisted ergonomic design can significantly enhance construction practices. For

   instance, in a simulated analysis of a hospital construction site, the proposed AI system identified excessive upper limb elevation during overhead HVAC installations. Based on this output, the design team modified ceiling access points and work heights, reducing predicted REBA scores from high to moderate risk levels [1] [4] [8] .

   1. Risk Assessment Requirements and Exposure Control This anticipatory approach makes it possible to document compliance with international standards such as ISO 11228,

      OSHA 1926, and local equivalents [6] [10]. AI-based predictive modules can serve as supplementary risk analysis tools, automatically identifying high-risk zones and suggesting corrective actions directly within digital design plans [1] [13].

   2. Maintenance and Decommissioning Procedures

      By incorporating ergonomic simulations early in the design process, safer and more accessible procedures can be pre-planned [1] [7]. This is especially valuable in industrial projects, hospitals, or large-scale infrastructure, where ongo- ing maintenance is intensive and critical.

   3. Training and Personal Protective Equipment (PPE)

      Early identification of ergonomic hazards also supports more accurate workforce training and better planning of personal protective equipment (PPE) needs. This allows training pro- grams to be tailored to actual ergonomic risks rather than generic scenarios [5] [9].

5. DISCUSSION

   The anticipated outcomes of the proposed framework suggest that the use of artificial intelligence (AI) during architectural design processes can predict potentially hazardous ergonomic conditions with greater precision than conventional post-hoc analysis methods [1] [4] [8] [13].

   This predictive capacity, enabled through collaborative digital modeling platforms such as BIM, facilitates early interven- tion, allowing spatial and functional decisions to be corrected before they materialize as harmful environments [7] [13].

   The comparison between AI-generated predictions and tradi- tional ergonomic tools (e.g., REBA, NIOSH) indicates AIs potential to complement or outperform manual evaluations [4] [8] [9].

   While standards such as ISO and OSHA provide valuable guidelines, they were not designed for digital design work- flows [6] [10]. This methodology helps push regulations toward a more proactive and integrated model.

   Further, dynamic, data-driven practices should be adopted using predictive analytics and integrating ergonomics as a core design variable from the start [1] [7].

   However, limitations include data scarcity, variability across countries, and regulatory fragmentation [11] [12]. Ethical con- siderations such as algorithmic accountability and human oversight are also essential.

   6.1. Study Limitations

   Despite the promising outcomes of this framework, several limitations must be acknowledged. First, the absence of empirical testing in real-world construction sites means that the algorithmic predictions have yet to be validated under dynamic field conditions. Second, the quality and representa- tiveness of the training data are constrained by the availability of ergonomic datasets, which may not fully capture regional variations, gender differences, or specific trade activities.

   Additionally, regulatory differences across countries may affect the scalability and standardization of the approach. These limitations highlight the need for further testing, broader data integration, and collaborative efforts to adapt the methodology to diverse construction conexts.

6. CONCLUSIONS AND FUTURE WORK

   This framework shows that bringing artificial intelligence into the early design stages of construction projects can shift the way we approach workplace safety. Instead of reacting to risks once workers are already exposed, this method helps identify and reduce ergonomic hazards from the very begin- ning when decisions about space, materials, and workflows are still flexible [1] [13].

   By simulating real-world tasks and environments, this ap- proach enables data-driven design decisions that not only help prevent musculoskeletal disorders (MSDs) but also enhance productivity and ensure regulatory compliance [4] [6] [10].

   Looking ahead, there are several areas that deserve deeper exploration:

   • Testing the system in real-life construction projects.

   • Expanding ergonomic datasets to reflect different popular

     ‌tions and work types.‌‌‌

     • Creating BIM plug-ins for real-time feedback during design.

     • Aligning the tool with international regulations

   and ensuring that ethical concerns like transparency and human oversight are built into every stage [1] [5] [7] [11].

   In conclusion, this research offers not only a technological innovation but a paradigm shift toward human-centered, prevention-oriented design in the construction sector.

   More than just a technical solution, this research contributes to a design culture that puts people first. It points toward a future where technology doesn't just build structures but also protects the health, dignity, and well-being of those who build them. Ultimately, integrating human-centered design and artificial intelligence at the earliest stages of development represents a transformative shift toward safer, smarter, and more sustainable construction environments.

7. REFERENCES

1. Abdelgawad A, Oti AH. Integration of ergonomics in BIM for con- struction safety. Automat Constr. 2022;134:104044. doi: 10.1016/j.autcon.2021.104044

2. Burry M. Scripting cultures: Architectural design and programming. Chichester: Wiley; 2011. (DOI no disponible, libro impreso)

3. Duarte JP. A discursive grammar for customizing mass housing: The case of Sizas houses at Malagueira. Automat Constr. 2005;14(2):265275. doi: 10.1016/j.autcon.2004.08.010

4. Hignett S, McAtamney L. Rapid Entire Body Assessment (REBA). Appl Ergon. 2000;31(2):201205. doi:

   10.1016/S0003-6870(99)00039-3

5. Hsiao H, Long D, Snyder K. Anthropometric procedures for sizing PPE. Hum Factors. 2002;44(1):114. doi: 10.1518/0018720024494855

6. International Organization for Standardization. ISO 11226:2000 Er- gonomics Evaluation of static working postures. Geneva: ISO; 2000. (DOI no disponible, norma técnica impresa)

7. Lu Q, Won J, Cheng JCP. A BIM- and IoT-based smart system for occupational health and safety management on construction sites. Int J Environ Res Public Health. 2016;13(11):1086.doi: 10.3390/ijerpp3111086

8. McAtamney L, Corlett EN. RULA: A survey method for the investiga- tion of work-related upper limb disorders. Appl Ergon. 1993;24(2):9199. doi: 10.1016/0003-6870(93)90080-S

9. National Institute for Occupational Safety and Health (NIOSH). Applications manual for the revised NIOSH lifting equation. Cincin- nati: U.S. Department of Health and Human Services; 1991. (DOI no disponible, documento gubernamental)

10. Occupational Safety and Health Administration (OSHA). Construction Industry Regulations (29 CFR Part 1926). Washington, DC: U.S. Department of Labor. Availablefrom:https://www.osha.gov/laws-regs/regulations/standardn umber/1926

11. Punnett L, Wegman DH. Work-related musculoskeletal disorders: the epidemiologic evidence and the debate. J Electromyogr Kinesiol. 2004;14(1):1323. doi: 10.1016/j.jelekin.2003.09.015

12. ‌Schneider S, Irastorza X. OSH in figures: Work-related musculoskel- etal disorders in the EU Facts and figures. Bilbao: European Agency for Safety and Health at Work; 2010. Available from: https://osha.europa.eu/en/publications/osh-figures-work-related-musc uloskeletal-disorders-eu-facts-and-figures

13. Zhang S, Teizer J, Lee JK, Eastman CM, Venugopal M. Building Information Modeling (BIM) and safety: Automatic safety checking of construction models and schedules. Automat Constr. 2013;29:183

195. doi: 10.1016/j.autcon.2012.05.006