Hasil untuk "Industrial safety. Industrial accident prevention"

工業講話会

Daniel Ruan, Salma Mozaffari, Sigrid Adriaenssens et al.

Industrial robots are increasingly deployed in contact-rich construction and manufacturing tasks that involve uncertainty and long-horizon execution. While learning-based visuomotor policies offer a promising alternative to open-loop control, their deployment on industrial platforms is challenged by a large observation-execution gap caused by sensing, inference, and control latency. This gap is significantly greater than on low-latency research robots due to high-level interfaces and slower closed-loop dynamics, making execution timing a critical system-level issue. This paper presents a latency-aware framework for deploying and evaluating visuomotor policies on industrial robotic arms under realistic timing constraints. The framework integrates calibrated multimodal sensing, temporally consistent synchronization, a unified communication pipeline, and a teleoperation interface for demonstration collection. Within this framework, we introduce a latency-aware execution strategy that schedules finite-horizon, policy-predicted action sequences based on temporal feasibility, enabling asynchronous inference and execution without modifying policy architectures or training. We evaluate the framework on a contact-rich industrial assembly task while systematically varying inference latency. Using identical policies and sensing pipelines, we compare latency-aware execution with blocking and naive asynchronous baselines. Results show that latency-aware execution maintains smooth motion, compliant contact behavior, and consistent task progression across a wide range of latencies while reducing idle time and avoiding instability observed in baseline methods. These findings highlight the importance of explicitly handling latency for reliable closed-loop deployment of visuomotor policies on industrial robots.

Qishen Ye, Yihai Fang, Nan Zheng

Wesley Tsz-Kin Chan, Wen-Chin Li, Richard Yeun et al.

Матвеев В.А., Ничкова Л.А.

Tuğçe Bilen, Mehmet Ozdem

The convergence of Information Technology (IT) and Operational Technology (OT) is a critical enabler for achieving autonomous and intelligent industrial systems. However, the increasing complexity, heterogeneity, and real-time demands of industrial environments render traditional rule-based or static management approaches insufficient. In this paper, we present a modular framework based on the Knowledge-Defined Networking (KDN) paradigm, enabling adaptive and autonomous control across IT-OT infrastructures. The proposed architecture is composed of four core modules: Telemetry Collector, Knowledge Builder, Decision Engine, and Control Enforcer. These modules operate in a closed control loop to continuously observe system behavior, extract contextual knowledge, evaluate control actions, and apply policy decisions across programmable industrial endpoints. A graph-based abstraction is used to represent system state, and a utility-optimization mechanism guides control decisions under dynamic conditions. The framework's performance is evaluated using three key metrics: decision latency, control effectiveness, and system stability, demonstrating its capability to enhance resilience, responsiveness, and operational efficiency in smart industrial networks.

SoYoung Park, Hyewon Lee, Mingyu Choi et al.

Anomaly segmentation is essential for industrial quality, maintenance, and stability. Existing text-guided zero-shot anomaly segmentation models are effective but rely on fixed prompts, limiting adaptability in diverse industrial scenarios. This highlights the need for flexible, context-aware prompting strategies. We propose Image-Aware Prompt Anomaly Segmentation (IAP-AS), which enhances anomaly segmentation by generating dynamic, context-aware prompts using an image tagging model and a large language model (LLM). IAP-AS extracts object attributes from images to generate context-aware prompts, improving adaptability and generalization in dynamic and unstructured industrial environments. In our experiments, IAP-AS improves the F1-max metric by up to 10%, demonstrating superior adaptability and generalization. It provides a scalable solution for anomaly segmentation across industries

Hai Dinh-Tuan

The convergence of Information Technology (IT) and Operational Technology (OT) in Industry 4.0 exposes the limitations of traditional, hierarchical architectures like ISA-95 and RAMI 4.0. Their inherent rigidity, data silos, and lack of support for cloud-native technologies impair the development of scalable and interoperable industrial systems. This paper addresses this issue by introducing CRACI, a Cloud-native Reference Architecture for the Industrial Compute Continuum. Among other features, CRACI promotes a decoupled and event-driven model to enable flexible, non-hierarchical data flows across the continuum. It embeds cross-cutting concerns as foundational pillars: Trust, Governance & Policy, Observability, and Lifecycle Management, ensuring quality attributes are core to the design. The proposed architecture is validated through a two-fold approach: (1) a comparative theoretical analysis against established standards, operational models, and academic proposals; and (2) a quantitative evaluation based on performance data from previously published real-world smart manufacturing implementations. The results demonstrate that CRACI provides a viable, state-of-the-art architecture that utilizes the compute continuum to overcome the structural limitations of legacy models and enable scalable, modern industrial systems.

Alpana Dubey, Suma Mani Kuriakose, Nitish Bhardwaj

We propose an approach to generate synthetic data to train computer vision (CV) models for industrial wear and tear detection. Wear and tear detection is an important CV problem for predictive maintenance tasks in any industry. However, data curation for training such models is expensive and time-consuming due to the unavailability of datasets for different wear and tear scenarios. Our approach employs a vision language model along with a 3D simulation and rendering engine to generate synthetic data for varying rust conditions. We evaluate our approach by training a CV model for rust detection using the generated dataset and tested the trained model on real images of rusted industrial objects. The model trained with the synthetic data generated by our approach, outperforms the other approaches with a mAP50 score of 0.87. The approach is customizable and can be easily extended to other industrial wear and tear detection scenarios

Bingyang Guo, Hongjie Li, Ruiyun Yu et al.

3D anomaly detection (3D-AD) plays a critical role in industrial manufacturing, particularly in ensuring the reliability and safety of core equipment components. Although existing 3D datasets like Real3D-AD and MVTec 3D-AD offer broad application support, they fall short in capturing the complexities and subtle defects found in real industrial environments. This limitation hampers precise anomaly detection research, especially for industrial equipment components (IEC) such as bearings, rings, and bolts. To address this challenge, we have developed a point cloud anomaly detection dataset (IEC3D-AD) specific to real industrial scenarios. This dataset is directly collected from actual production lines, ensuring high fidelity and relevance. Compared to existing datasets, IEC3D-AD features significantly improved point cloud resolution and defect annotation granularity, facilitating more demanding anomaly detection tasks. Furthermore, inspired by generative 2D-AD methods, we introduce a novel 3D-AD paradigm (GMANet) on IEC3D-AD. This paradigm generates synthetic point cloud samples based on geometric morphological analysis, then reduces the margin and increases the overlap between normal and abnormal point-level features through spatial discrepancy optimization. Extensive experiments demonstrate the effectiveness of our method on both IEC3D-AD and other datasets.

Ok-seol Kim, Gil-sang Jang, Seok-in Mun

Pankaj Kumar Gupta, Manish Kumar Singh

This research presents an IoT-based intelligent safety system designed to monitor and prevent multiple hazards, including fire, gas leaks, and industrial accidents. The system integrates temperature, smoke, gas, vibration, and current sensors with an ESP32 microcontroller and NB-IoT communication modules. Data is processed and analysed using AWS IoT cloud services, enabling real-time monitoring and timely alerts through mobile notifications and audible alarms. Experimental results show high detection accuracy and reduced response time, making it suitable for industrial plants, smart buildings, and public safety applications.

Hyeon Jo, E. Baek

Occupational noise exposure is a pervasive issue in many industries, leading to a range of health issues and sleep disturbances among workers. Additionally, there is a strong desire among these workers to prevent industrial accidents. This study, aimed at enhancing worker health and well-being, utilized a survey distributed by the Korean Confederation of Trade Unions to field workers. Data from 1285 workers were collected and analyzed using partial least squares structural equation modeling (PLS-SEM) to identify and understand the factors affecting prevention intention in noisy work environments. Our findings indicate that health problems resulting from occupational noise exposure significantly influence insomnia, perceived severity of potential accidents, perceived benefits of preventive measures, and perceived barriers. Perceived severity was significantly correlated with prevention intention, emphasizing the role of risk perception in motivating preventive behaviors. Perceived benefits were also significantly associated with prevention intention, highlighting the importance of positive outcomes in influencing workers’ behaviors. Additionally, perceived barriers showed a significant relationship with prevention intention, suggesting that overcoming these barriers is crucial in promoting preventive behaviors. Demographic factors such as gender displayed a significant association with prevention intention, while age did not. This study provides valuable insights into the multifaceted factors influencing workers’ intention to prevent industrial accidents in noisy environments, underlining the importance of comprehensive data collection tools in understanding these dynamics.

Siti Saroh Tanwir, Ahmad Syaiful Huda, Abdul Latif et al.

Work accidents are not just a single event but occur through a series of interrelated causes. The main source of accidents is the existence of unsafe actions that refer to worker behavior and unsafe conditions that refer to the work environment. The approach that can be used to prevent work accidents in the work environment is the implementation of the 5S standard through the Plan Do Check Action (PDCA) methodology. The implementation of 5S is carried out in stages and systematically starting from planning, implementation, checking, and continual improvement of each 5S element. Each of the five stages is important and should be taken seriously and handled separately and sequentially. The initial three stages are operational; the fourth phase preserves the state established in the first three phases; and the fifth phase aids in our efforts to improve things continuously. Improved working conditions and an accident-free workplace can result from a better understanding of the 5S idea and how it relates to the safety management system. This will increase employee satisfaction in industrial organizations.

Putri Ayuni Alayyannur, Muhammad Malik Al Hakim, Rr. Sri Rejeki Eviyanti Puspita Sari

Introduction: Every workplace must make an occupational health effort to avoid health problems. Many workers underestimate the risks of the job and, therefore, do not use safety equipment even when available. The most often reported case of occupational skin illnesses, contact dermatitis, accounts for more than 85% of all cases. This study was conducted to occupational dermatitis and its relationship to personal protective equipment (PPE) use. Methods: The literature search was carried out in April 2021. The research sources were taken from several databases with the keywords dermatitis, occupational health, and personal protective equipment. The Google Scholar database found 17,710 articles, ScienceDirect found 1,264 articles, ProQuest found 888 articles, and PubMed found 452 articles. Of the entire database, only 36 articles met the inclusion criteria. Results: This literature review shows that dermatitis is experienced by workers in various sectors including health workers, hairdressers, scavengers, farmers, fishermen, manufacturing industry workers, printing workers, and construction workers. The use of PPE can reduce the risk of dermatitis. However, in some conditions, the use of PPE has no effect or can even cause dermatitis due to irritation and allergies to the ingredients contained in the PPE. The limitation of this research is that the articles that are the source of this review are only from 2016–2021.Conclusion: Dermatitis still occurs in various occupational sectors. The risk of dermatitis can be decreased by using PPE; however, it can also cause the occurrence of dermatitis itself.

Ashok Urlana, Charaka Vinayak Kumar, Ajeet Kumar Singh et al.

Large language models (LLMs) have become the secret ingredient driving numerous industrial applications, showcasing their remarkable versatility across a diverse spectrum of tasks. From natural language processing and sentiment analysis to content generation and personalized recommendations, their unparalleled adaptability has facilitated widespread adoption across industries. This transformative shift driven by LLMs underscores the need to explore the underlying associated challenges and avenues for enhancement in their utilization. In this paper, our objective is to unravel and evaluate the obstacles and opportunities inherent in leveraging LLMs within an industrial context. To this end, we conduct a survey involving a group of industry practitioners, develop four research questions derived from the insights gathered, and examine 68 industry papers to address these questions and derive meaningful conclusions. We maintain the Github repository with the most recent papers in the field.

Xuejiao Li, Cheng Yang, Charles Møller et al.

In the era of Industry 4.0, artificial intelligence (AI) is assuming an increasingly pivotal role within industrial systems. Despite the recent trend within various industries to adopt AI, the actual adoption of AI is not as developed as perceived. A significant factor contributing to this lag is the data issues in AI implementation. How to address these data issues stands as a significant concern confronting both industry and academia. To address data issues, the first step involves mapping out these issues. Therefore, this study conducts a meta-review to explore data issues and methods within the implementation of industrial AI. Seventy-two data issues are identified and categorized into various stages of the data lifecycle, including data source and collection, data access and storage, data integration and interoperation, data pre-processing, data processing, data security and privacy, and AI technology adoption. Subsequently, the study analyzes the data requirements of various AI algorithms. Building on the aforementioned analyses, it proposes a data management framework, addressing how data issues can be systematically resolved at every stage of the data lifecycle. Finally, the study highlights future research directions. In doing so, this study enriches the existing body of knowledge and provides guidelines for professionals navigating the complex landscape of achieving data usability and usefulness in industrial AI.

Fei-Fei Ye, Long-Hao Yang, Ying-Ming Wang et al.

Ricardo P. Arciniega-Rocha, Vanessa C. Erazo-Chamorro, Gyula Szabo

The industrial sector is improving its management systems and designing healthy workspaces by focusing on selecting the best ways to reduce accidents and optimize financial and human resources. Hand tools represent the general equipment used in a significant range of industrial jobs. This research aims to develop a tool selection method to help users, managers, and tool designers ensure awareness and care regarding ergonomics based on the anthropometrics of employees, considering the main risk factors for tool selection. The information, which relates to hand security risk factors and the established parameters set by official international institutions, is evaluated during the study. This paper also presents a safety risk assessment framework based on criteria collected through a survey from 10 experts to rate the initial risk value and determine its importance using the analytical hierarchy process (AHP). As a result, the analysis identified the possibility of injury (with 73.06% accuracy) as the biggest concern for companies due to its immediate effects on workers’ health. It provides a decision regimen—a tool for decision-makers to design and plan prevention activities to reduce accidents, injuries, and possible illnesses. It further lays out a methodical and analytical model to be used by managers to ensure correct hand tool selection. This model can be used to reduce the possibility of illnesses or injuries for workers and tailor the ergonomic design of each workstation according to specific hand anthropometric data for the worker.

Sandra Lange, Wioletta Mędrzycka-Dąbrowska, Anna Małecka-Dubiela

Introductions: Computed tomography is one of the biggest breakthroughs in diagnostic imaging. In order to more accurately assess structures and pathological changes during the examination, it is necessary to administer a contrast agent. Patients presenting for the examination, very often only find out during the examination that a contrast agent is required. This increases patients’ uncertainty when giving written consent for contrast administration, as well as anxiety about the examination. The aim of this study was to explore the experiences of patients who have contrast-enhanced CT scans, focusing primarily on anxiety, feelings, and safety. Methods: The cross-sectional study was conducted in diagnostic imaging offices in Pomeranian Voivodeship in 2019–2020. The survey was aimed at patients presenting for CT examinations with intravenous contrast injection. In total, 172 patients participated in the survey. A proprietary survey questionnaire was used to conduct the study. Results and Conclusions: During a CT scan, intravenous contrast agent administration is often necessary. Although there are few studies on patients’ experiences with this examination, the authors observe that some patients experience anxiety. The results of our study showed the following: (1) 30.8% of patients experience anxiety before a CT scan with intravenous contrast injection; (2) variables such as gender, previous experience, and searching for information about the examination influence the occurrence of anxiety; (3) the most common feelings experienced by patients during intravenous contrast injection are a sensation of warmth spreading throughout the body; (4) the most common source of information about the study used among patients was the Internet; (5) most patients feel safe during a CT scan.