International Journal of Industrial Ergonomics最新文献

{"title":"A multimodal physical fatigue assessment method using a biomarker and accelerometer-embedded wearable wristband","authors":"Md Hadisur Rahman,&nbsp;JuHyeong Ryu","doi":"10.1016/j.ergon.2025.103867","DOIUrl":"10.1016/j.ergon.2025.103867","url":null,"abstract":"<div><div>Physically demanding tasks pose significant challenges to worker health, safety, and productivity across various industrial sectors, particularly in construction. Physical fatigue, a major contributor to workplace accidents, compromises individual well-being and economic outcomes. Traditional fatigue assessment methods often lack accuracy, comfort, or real-world applicability. This study introduces a multimodal physical fatigue assessment method employing a wearable sensor to collect both physiological data, heart rate (HR) and skin temperature (ST), and kinematic data, jerk. The Borg's Rating of Perceived Exertion (RPE) scale was used to validate the method, which was tested on twenty-two participants (mean age: 28.5 ± 3.6 years) performing manual material handling tasks. The findings indicate that HR, ST, and jerk values increase as tasks become more strenuous, correlating with higher RPE scores. Integrating physiological and kinematic metrics with subjective validation effectively captures the multifaceted nature of fatigue, enabling real-time monitoring. Notably, incorporating jerk as a kinematic measure addresses limitations of previous methods by providing a rapid-response indicator of motor control and ultimately physical fatigue. This comprehensive approach has potential applications in manual material handling tasks and, with further validation, may be extended to other industrial contexts where repetitive lifting and carrying are common. By offering practical, data-driven solutions that enhance workplace safety and health, this approach can reduce accidents, injuries and support proactive risk management strategies.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103867"},"PeriodicalIF":3.0,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Work and women's sacral spine acute injuries: an underestimated risk","authors":"Claudia Giliberti ,&nbsp;Silvana Salerno","doi":"10.1016/j.ergon.2025.103869","DOIUrl":"10.1016/j.ergon.2025.103869","url":null,"abstract":"<div><h3>Introduction</h3><div>Spine is the third most commonly injured anatomical region among Italian working women, after upper and lower limbs. Lumbar spine work injuries are well-known, particularly in the Healthcare sector, while no studies on sacral spine work-related injuries were found, although they represent a “silent epidemic” for general population, producing severe disabilities among women.</div></div><div><h3>Objective</h3><div>The aim of this study is to analyze sex/gender differences in compensated work-related sacral spine injuries in mainly female-dominated work sectors.</div></div><div><h3>Methods</h3><div>Compensated work-related acute sacral spine injuries among women and men<strong><u>,</u></strong> from the Italian Compensation Authority (Inail) database in the last five years, were studied in selected work sectors and the statistical analysis was performed as Incidence Rate Ratio (IRR) and Odds Ratio (OR) (p &lt; 0,05). Sacral spine work-related lesions such as bruises, dislocations and fractures were analyzed per sex/gender and work sectors.</div></div><div><h3>Results</h3><div>Women showed a statistically significant IRR for sacral spine work-related injuries in all the analyzed work sectors (IRR 2.22; CI95 % 2.11–2.34), especially in Catering, Cleaning and Trade. Women suffered more sacral fractures than men (OR 1.28; CI95 % 1.14–1.44), especially in Manufacturing (OR 1.47; CI95 % 1.08–1.99), where women are mainly employed in food processing. The role of work falls is discussed, together with the need of an intersectional ergonomic approach to prevent this underestimated risk among women.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103869"},"PeriodicalIF":3.0,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Guiding by touch: A vibrotactile navigation system for underwater situational awareness","authors":"Giovanna Camacho ,&nbsp;Matthew L. Bolton ,&nbsp;Amanda Watson ,&nbsp;Henry Bearden ,&nbsp;Sharon Lu","doi":"10.1016/j.ergon.2025.103864","DOIUrl":"10.1016/j.ergon.2025.103864","url":null,"abstract":"<div><div>This study will test the usability of wearable, vibrotactile cues in providing intuitive orientation and communication cues to participants in visually challenging underwater navigation tasks. The device’s signals were designed to communicate the three levels of situational awareness (SA; perceive, comprehend, and project) intuitively, as if one was being guided by a partner’s hand. We evaluated the effectiveness of this device in a human subject experiment with divers wearing fully blacked-out dive masks. Performance with the vibrotactile display was compared against the Scubapro heads-up display, along with dive rescue team rope pulls based on performance measures (navigation, accuracy, and time). Subjective measures of mental workload, situational awareness, and usability were collected; as well as surveys designed to understand how participants classified tactor signals into SA levels. The results showed that the tactile design enhanced accuracy, but increased navigation time. This design was comparable to other standard methods across subjective mental workload, SA, and usability measures. The paper discusses the significance of these results for the navigation of both commercial and professional divers. It also explores the implications for navigation support in other visually challenging environments.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103864"},"PeriodicalIF":3.0,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145798266","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Workstation ergonomics in the era of multi-monitor technology: A narrative review and survey","authors":"Andrea A. Vivaldi,&nbsp;David Claudio,&nbsp;Maria A. Velazquez,&nbsp;Laura Punnett","doi":"10.1016/j.ergon.2025.103863","DOIUrl":"10.1016/j.ergon.2025.103863","url":null,"abstract":"<div><div>Multi-monitor workstations are becoming more common, offering productivity gains and better workflow. However, their ergonomic impact is still not well understood, and current guidelines have not kept up with technology. This review looks at published studies on multi-monitor setups and finds mixed methods and results on productivity, user comfort, and musculoskeletal risk. Differences in study design and reporting make it hard to reach clear conclusions, leaving gaps in guidance. To add real-world data, we surveyed 208 computer users (85 % confidence level, ±5 % margin of error). We used descriptive statistics and K-means clustering to explore patterns. About 65 % of respondents used multiple monitors. Out of 135 participants using multiple monitors, two-monitor setups were most common (71 %), with L-shaped layouts used by half of multi-monitor users. Cluster analysis showed four main user types, from triple-monitor “power users” (about 8.5 h/day) to laptop-focused dual-monitor users (about 7.2 h/day). These groups differed in screen size, layout, and how time was split between monitors. Current research lacks consistency and does not address newer options like ultrawide monitors, making practical guidance difficult. Survey data further reveal a growing reliance on dual-monitor configurations, with users likely adopting suboptimal arrangements, which may contribute to musculoskeletal discomfort. This study highlights the urgency of developing updated ergonomic recommendations and research that balance efficiency with user well-being, ensuring that productivity gains do not come at the cost of discomfort or injury.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103863"},"PeriodicalIF":3.0,"publicationDate":"2025-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145798267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessing pilot cognitive overload risk with a random forest framework: A non-contact approach based on a novel cardiopulmonary feature","authors":"Gu Sen ,&nbsp;Hou Wenjun ,&nbsp;Wang Hanyu ,&nbsp;Wang Qingbin","doi":"10.1016/j.ergon.2025.103865","DOIUrl":"10.1016/j.ergon.2025.103865","url":null,"abstract":"&lt;div&gt;&lt;div&gt;Accurate, real-time assessment of operator mental workload(MWL) is critical for ensuring safety and efficiency in complex systems like aviation. However, existing methods are limited by the latency of subjective scales and the intrusiveness of contact-based physiological sensors. In order to address the challenge of real-time, interference-free monitoring of operator cognitive states in complex systems such as flight, this study proposes and validates a new non-contact psychological load assessment method based on cardiopulmonary coupling joint entropy. We recorded electrocardiogram (ECG) and respiration (RSP) signals from thirty flight trainees using a 77 GHz millimeter-wave radar as they performed a simulated flight task with four escalating levels of difficulty. Subjective workload ratings (NASA-TLX) and task performance were collected concurrently for validation. Results confirmed the experimental manipulation's effectiveness, with NASA-TLX scores(&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; (3,87) = 34.87, &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt;0.001), reaction times(&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; (3,87) = 25.712, &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt; 0.001), and error rates(&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; (3,87) = 34.881, &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; &lt; 0.05) all increasing significantly with task difficulty. The joint entropy value exhibited a monotonic increase with workload levels(&lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; (3,87) = 16.578, &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; = 0.002), demonstrating its high sensitivity. Feature importance analysis identified joint entropy as the most significant predictor of MWL. Notably, a classification model utilising only the joint entropy feature achieved superior predictive accuracy compared to a model using the full feature set, highlighting the metric's robustness and efficiency. This study validates non-contact cardiopulmonary coupling joint entropy as a sensitive and reliable biomarker for MWL. This method provides a practical path for developing intelligent safety management systems capable of warning of cognitive overload, preventing human errors, and promoting adaptive human-machine collaboration. These systems can serve as key inputs for AI-driven adaptive interfaces, promoting dynamic human-machine collaboration in line with Industry 5.0 principles.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Relevance to industry&lt;/h3&gt;&lt;div&gt;This study presents a practical approach to the real-time, objective and non-intrusive monitoring of operators' MWL in high-risk sectors such as aviation and nuclear power. The verified joint entropy index can be integrated into safety management systems to develop AI-powered intelligent assistance systems and adaptive human-machine interfaces that can dynamically adjust to the operator's cognitive state. These systems can warn of cognitive overload, effectively preven","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103865"},"PeriodicalIF":3.0,"publicationDate":"2025-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145798268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Teatime Tales: A Deep Dive into the social sustainability of the tea garden ecosystem","authors":"Ankit Basak ,&nbsp;Shiv Kumar Verma","doi":"10.1016/j.ergon.2025.103841","DOIUrl":"10.1016/j.ergon.2025.103841","url":null,"abstract":"<div><div>Tea gardens are one of India's oldest organised sectors. However, people employed in the sector still face issues similar to those in the unorganised sector. Most studies in this domain have primarily focused on tea leaf production and operational efficiency rather than the social sustainability aspects of the tea garden ecosystem. Moreover, there is also a dearth of research that includes the perspective of females working in the domain. The current exploratory study examines the tea garden ecosystem to understand the various factors that affect workers and their working conditions, with a particular focus on gender dynamics and social sustainability. We adopted a qualitative research methodology for the study. Direct observation was conducted to study the behaviour of workers, various processes, and events in the tea garden. Videos and images were collected for a visual ethnography. Lastly, we conducted semi-structured interviews with various stakeholders in the tea garden, including workers, administrators, managers, and doctors, followed by thematic analysis to analyse the collected data. The findings reveal persistent challenges such as low wages, gender-based division of labour, lack of ergonomic support, and even the influence of gender-specific clothing on workers' health, particularly during pesticide spraying. The findings from direct observation, visual ethnography, and semi-structured interviews were then combined to provide a comprehensive view of the various issues, challenges, and working conditions faced by the tea garden workers. The research provides a foundational understanding that can inform policy, design, and future interdisciplinary studies aimed at enhancing social sustainability in similar labour-intensive sectors.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103841"},"PeriodicalIF":3.0,"publicationDate":"2025-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145749070","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel driver monitoring and feedback system improves takeover performance in conditional automated driving","authors":"N. Zhang ,&nbsp;M. Fard ,&nbsp;S. Tohmuang ,&nbsp;J. Xu ,&nbsp;J.L. Davy ,&nbsp;S.R. Robinson","doi":"10.1016/j.ergon.2025.103862","DOIUrl":"10.1016/j.ergon.2025.103862","url":null,"abstract":"<div><div>As automated driving evolves, ensuring seamless human-vehicle interaction remains a critical challenge. Building upon authors’ previous study, the present study develops and investigates how a Driver Monitoring and Feedback System (DMFS) influences takeover performance, physiological responses, and user experience in conditional automated driving. Seventeen participants engaged in simulated driving sessions with two Non-Driving Related Tasks (NDRTs), namely working and resting, both with and without an active DMFS. Metrics were collected for driving performance, Heart Rate Variability (HRV) and subjective evaluations. The findings indicate that the DMFS mitigated the adverse effects of NDRTs on takeover performance by up to 47 %, particularly during resting conditions. Although the DMFS was generally perceived positively regarding effectiveness and accuracy, lower user experience scores suggest a need for a balance between functionality and user comfort. This study highlights the potential of DMFSs to enhance safety in automated driving, while also identifying challenges in maintaining driver readiness and optimising human-automation interaction. The results underscore the importance of developing adaptive, user-centric DMFS designs for future automated driving systems.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103862"},"PeriodicalIF":3.0,"publicationDate":"2025-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145749069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hand function ergonomic determinants in aging: Effects of sex, age-group, and muscle mass in Taiwanese older adults","authors":"Yi-Lang Chen ,&nbsp;Syuan Yu ,&nbsp;Yu-Chi Lee","doi":"10.1016/j.ergon.2025.103852","DOIUrl":"10.1016/j.ergon.2025.103852","url":null,"abstract":"<div><div>As Taiwan enters a super-aged era, understanding the physiological determinants of hand function is essential for promoting independence and quality of life in older adults. This study examined the effects of sex, age group (&lt;70 vs. ≥70 years), and testing posture on grip strength, pinch strength, and wrist range of motion (ROM), and explored their associations with segmental body composition. One hundred right-handed, community-dwelling older adults (50 males, 50 females) underwent standardized assessments. Grip strength was significantly affected by sex and age (both <em>p</em> &lt; 0.001), with males outperforming females across all postures and a marked age-related decline observed in women aged ≥70. Testing posture did not significantly influence grip strength. Pinch strength differed by pinch type (<em>p</em> &lt; 0.001), with lateral pinch producing the highest values; a significant sex × pinch type interaction (<em>p</em> &lt; 0.001) showed the greatest male–female disparity in lateral pinch. Wrist ROM was generally preserved, although a modest sex × age interaction for radial deviation (<em>p</em> &lt; 0.01) was found. Skeletal muscle mass strongly predicted grip strength (R² = 0.525–0.550), while trunk and upper-limb muscle mass differentially predicted pinch strength (R² = 0.330–0.557). These findings emphasize the need for age- and sex-specific ergonomic standards and muscle-targeted interventions in aging populations.</div></div><div><h3>Relevance to industry</h3><div>This study offers ergonomic insights for designing tools, interfaces, and rehabilitation programs suited to aging populations. By identifying sex-specific strength differences and muscle based hand function predictors, it supports the development of age-appropriate, task-specific interventions to maintain functional capacity and promote independence among older adults in daily life.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103852"},"PeriodicalIF":3.0,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145694794","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Detection of perceived risk during partially automated driving on real road","authors":"Yanyi Li ,&nbsp;Yi Wang ,&nbsp;Xin Zhou ,&nbsp;Wei Zhang ,&nbsp;Jingyue Zheng","doi":"10.1016/j.ergon.2025.103842","DOIUrl":"10.1016/j.ergon.2025.103842","url":null,"abstract":"<div><div>With the rapid advancement of automated driving technologies, understanding drivers' perceived risk in real-road environments is crucial for the acceptance of automated vehicles (AVs) and ensuring safety. This study proposed a deep learning framework combining convolutional neural network (CNN) and long short-term memory (LSTM) to detect the perceived risk. The framework utilized vehicle, environmental, physiological, and facial data collected from real-road experiments involving partially automated vehicles. Potential high perceived risk events were captured via a smartphone. They were further classified into low/medium/high levels through subjective evaluation. Compared with other methods, the CNN-LSTM model performs the best, achieving an accuracy of 82.8 %, an F1 score of 83.6 %, and an AUC score of 86.5 %. Key features sensitive to perceived risk changes include the steering angle, motorcycle count, skin conductance level, root mean square of electromyographic signals, and eye-related features. The model and the findings of the study may contribute to improved the design of take-over request and driving style adjustment for automated vehicles to improve safety and user acceptance.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103842"},"PeriodicalIF":3.0,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intelligent prediction of ergonomics evaluation metrics in human-AI collaboration based on machine learning","authors":"Jiabing Zhang,&nbsp;Qingxuan Jia,&nbsp;Siyi Li,&nbsp;Shiyu Zhang,&nbsp;Gang Chen","doi":"10.1016/j.ergon.2025.103851","DOIUrl":"10.1016/j.ergon.2025.103851","url":null,"abstract":"<div><div>Ergonomic evaluations in human-AI collaboration systems are often time-consuming, labor-intensive, and prone to bias. The complexity of formulas and factors complicates automated ergonomic evaluation. To address this issue, this paper proposes a machine learning-based framework for predicting ergonomics evaluation metrics. First, a human-AI collaboration experimental method is presented for ergonomics evaluation metric data collection. During the experiment, a total of 18 human-AI collaborative experiments were conducted, comprising 12 human-AI teams and 6 all-human teams, covering various tasks like Flight, Attack, etc. Then, four regression models-linear neural network, nonlinear deep neural network, random forest regressor with data augmentation, and random forest with k-fold cross-validation-are designed to predict ergonomics evaluation metrics. Data augmentation techniques are employed to expand the dataset and enhance the model’s generalization capability. The dataset grew from 72 to 72,000 samples for neural networks and 7200 for random forests through data augmentation. The random forest model was also trained with 6-fold cross-validation, and test sets for all models were derived from the original data. Finally, the models’ accuracy and reliability in predicting these metrics are comprehensively evaluated using relative absolute error, mean relative absolute error, root mean squared error, relative root mean squared error, and the coefficient of determination <span><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span>, ensuring their validity. The results show that the proposed machine learning framework achieves high prediction accuracy, especially the data-augmented random forest model, which outperforms other models in terms of prediction accuracy. This enables effective automated ergonomic evaluations within human-AI collaboration systems and thus guides designing fluent and efficient human-AI teams.</div></div>","PeriodicalId":50317,"journal":{"name":"International Journal of Industrial Ergonomics","volume":"111 ","pages":"Article 103851"},"PeriodicalIF":3.0,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145658445","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}