Complex Industrial Environments Research Articles

Augmented Reality in Industry 4.0 Assistance and Training Areas: A Systematic Literature Review and Bibliometric Analysis

Augmented reality (AR) technology is making a strong appearance on the industrial landscape, driven by significant advances in technological tools and developments. Its application in areas such as training and assistance has attracted the attention of the research community, which sees AR as an opportunity to provide operators with a more visual, immersive and interactive environment. This article deals with an analysis of the integration of AR in the context of the fourth industrial revolution, commonly referred to as Industry 4.0. Starting with a systematic review, 60 relevant studies were identified from the Scopus and Web of Science databases. These findings were used to build bibliometric networks, providing a broad perspective on AR applications in training and assistance in the context of Industry 4.0. The article presents the current landscape, existing challenges and future directions of AR research applied to industrial training and assistance based on a systematic literature review and citation network analysis. The findings highlight a growing trend in AR research, with a particular focus on addressing and overcoming the challenges associated with its implementation in complex industrial environments.

Text Detection and Recognition for X-ray Weld Seam Images

X-ray weld seam images carry vital information about welds. Leveraging graphic–text recognition technology enables intelligent data collection in complex industrial environments, promising significant improvements in work efficiency. This study focuses on using deep learning methods to enhance the accuracy and efficiency of detecting weld seam information. We began by actively gathering a dataset of X-ray weld seam images for model training and evaluation. The study comprises two main components: text detection and text recognition. For text detection, we employed a model based on the DBNet algorithm and tailored post-processing techniques to the unique features of weld seam images. Through model training, we achieved efficient detection of the text regions, with 91% precision, 92.4% recall, and a 91.7% F1 score on the test dataset. In the text recognition phase, we introduced modules like CA, CBAM, and HFA to capture the character position information and global text features effectively. This optimization led to a remarkable text line recognition accuracy of 93.4%. In conclusion, our study provides an efficient deep learning solution for text detection and recognition in X-ray weld seam images, offering robust support for weld seam information collection in industrial manufacturing.

Enhanced YOLOv5s-Based Algorithm for Industrial Part Detection.

In complex industrial environments, accurate recognition and localization of industrial targets are crucial. This study aims to improve the precision and accuracy of object detection in industrial scenarios by effectively fusing feature information at different scales and levels, and introducing edge detection head algorithms and attention mechanisms. We propose an improved YOLOv5-based algorithm for industrial object detection. Our improved algorithm incorporates the Crossing Bidirectional Feature Pyramid (CBiFPN), effectively addressing the information loss issue in multi-scale and multi-level feature fusion. Therefore, our method can enhance detection performance for objects of varying sizes. Concurrently, we have integrated the attention mechanism (C3_CA) into YOLOv5s to augment feature expression capabilities. Furthermore, we introduce the Edge Detection Head (EDH) method, which is adept at tackling detection challenges in scenes with occluded objects and cluttered backgrounds by merging edge information and amplifying it within the features. Experiments conducted on the modified ITODD dataset demonstrate that the original YOLOv5s algorithm achieves 82.11% and 60.98% on mAP@0.5 and mAP@0.5:0.95, respectively, with its precision and recall being 86.8% and 74.75%, respectively. The performance of the modified YOLOv5s algorithm on mAP@0.5 and mAP@0.5:0.95 has been improved by 1.23% and 1.44%, respectively, and the precision and recall have been enhanced by 3.68% and 1.06%, respectively. The results show that our method significantly boosts the accuracy and robustness of industrial target recognition and localization.

IIoT Low-Cost ZigBee-Based WSN Implementation for Enhanced Production Efficiency in a Solar Protection Curtains Manufacturing Workshop.

Nowadays, the Industry 4.0 concept and the Industrial Internet of Things (IIoT) are considered essential for the implementation of automated manufacturing processes across various industrial settings. In this regard, wireless sensor networks (WSN) are crucial due to their inherent mobility, easy deployment and maintenance, scalability, and low power consumption, among other benefits. In this context, the presented paper proposes an optimized and low-cost WSN based on ZigBee communication technology for the monitoring of a real manufacturing facility. The company designs and manufactures solar protection curtains and aims to integrate the deployed WSN into the Enterprise Resource Planning (ERP) system in order to optimize their production processes and enhance production efficiency and cost estimation capabilities. To achieve this, radio propagation measurements and 3D ray launching simulations were conducted to characterize the wireless channel behavior and facilitate the development of an optimized WSN system that can operate in the complex industrial environment presented and validated through on-site wireless channel measurements, as well as interference analysis. Then, a low-cost WSN was implemented and deployed to acquire real-time data from different machinery and workstations, which will be integrated into the ERP system. Multiple data streams have been collected and processed from the shop floor of the factory by means of the prototype wireless nodes implemented. This integration will enable the company to optimize its production processes, fabricate products more efficiently, and enhance its cost estimation capabilities. Moreover, the proposed system provides a scalable platform, enabling the integration of new sensors as well as information processing capabilities.

Online human motion analysis in industrial context: A review

Human motion analysis plays a crucial role in industry 4.0 and, more recently, in industry 5.0 where human-centered applications are becoming increasingly important, demonstrating its potential for enhancing safety, ergonomics and productivity. Considering this opportunity, an increasing number of studies are proposing works on the analysis of human motion in an industrial context, taking advantage of the rise of artificial intelligence technologies and sensor technologies. The objective of this work is to provide a review of recent studies exploring these technologies in the analysis of human movement while specifically considering industrial context. First, a taxonomy of key human motion analysis applications is proposed, presenting statistical insights to reveal trends and highlighting lacks in current research. Furthermore, this work identifies benchmark datasets acquired in various industrial case studies and associated sensors. Many recommendations for selecting optimal sensors and valuable benchmarks are proposed. Then, the paper outlines the current trend of utilizing hybrid deep learning methodologies in human movement analysis while underscoring the performance and limitations of these proposed methods, considering industrial constraints such as real-time recognition and frugality. Finally, challenges and future works are highlighted, focusing on the opportunities to address problems related to the complex industrial environment in order to achieve reliable performances.

A high-precision crown control strategy for hot-rolled electric steel using theoretical model-guided BO-CNN-BiLSTM framework

The prediction accuracy of strip crown is low under complex industrial data environments to general machine learning models, i.e., lack of reasonable mechanism explanation and spatial dimension dependence, which will directly affect the product quality of hot-rolled electrical steel. Therefore, a high-precision crown prediction model is proposed for electrical steel in hot rolling based on a theoretical model-guided BO-CNN-BiLSTM (Bayesian optimization, Convolution Neural Network, and Bidirectional Long Short-term Memory) framework. The work roll wear model, thermal crown model, and the secondary deformation model of the strip between stands, based on the primary deformation by the loaded gap profile, were constructed. The mechanism parameters and measured parameters are integrated into a dataset as input feature variables. In the TG-BO-CNN-BiLSTM framework, the CNN-BiLSTM model, which can achieve spatial dimension dependence, was used to extract its feature component and sequentially predict the crown using the dataset, simultaneously, the BO module optimizes the hyperparameter of the CNN-BiLSTM model. The advantages of the proposed model are verified by adopting multiple evaluation indicators, which improves running speed and prediction accuracy. The effects of process parameters on the crown with typical upstream and downstream stands were comprehensively analyzed with the proposed model. A high-precision crown control strategy, combining the framework and influence law, for multi-stand and multi-method was proposed to obtain the high-precision crown. The control strategy applied in 1450 mm 4-high hot strip mills shows that the production performance of electrical steel is significantly improved.

Research on Visual Recognition and Localization Methods for Different Feature Parts

Aimed at the problems of complex industrial site environment, difficult identification of assembly features and low positioning accuracy, a new visual identification and positioning method is proposed, which can well identify and locate parts and products with different characteristics. First, combined filtering is used in preprocessing to repair low-quality images. Then, edge detection and contour finding are performed on the preprocessed image. Finally, the internal and external parameters of the camera are calculated through the camera calibration, and the acquired pixel coordinates are converted into world coordinates. The results show that the combined filtering algorithm has good noise reduction effect and can remove image surface noise, which improves the recognition accuracy for the identification and positioning of different feature parts by the subsequent contour search method, and the algorithm can meet the requirements of the actual positioning accuracy of the assembly.

Deep Reinforcement Learning with Inverse Jacobian based Model-Free Path Planning for Deburring in Complex Industrial Environment

Deep Reinforcement Learning with Inverse Jacobian based Model-Free Path Planning for Deburring in Complex Industrial Environment

Funding The Student Creativity Programs (PKM) In The Triple Helix For The Independence Of The Defense Industry

The urgency of funding student creativity programs within the Triple Helix framework to encourage the independence of the defense industry is a step forward in making economic progress, especially the defense industry. The SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis method is used as a research framework to evaluate the internal and external factors that influence the success of the program. This research reveals that funding for student creativity programs has a significant urgency in strengthening the defense industry. Through collaboration between students, industry and the government, this program is able to generate innovation, research and development of new technologies that can increase the competitiveness of the defense industry. The SWOT analysis shows that the strengths of the student creativity program in Triple Helix include access to fresh and innovative student intellectual resources, the ability to combine academic knowledge with the needs of the defense industry, and the encouragement of collaboration and knowledge exchange between all parties involved. Weaknesses identified include a lack of adequate funding, challenges in integrating student activities into a complex industrial environment, and a lack of understanding of the long-term benefits of the program. However, the SWOT analysis also identified significant opportunities, such as the increasing need for innovation and new technologies in the defense industry, government support for collaborative programs involving students, and the potential to develop long-term partnerships between the defense industry and universities. The threats faced include intense global competition, the need for constant updating of defense technology, and the potential for gaps between academic knowledge and industry needs if there is not adequate collaboration. In order to realize the independence of the defense industry, this research recommends adequate funding for student creativity programs, provision of supporting infrastructure, increased collaboration between universities, industry and government, as well as the development of a long-term strategy to ensure the sustainability of this program.

Heat-vision based drone surveillance augmented by deep learning for critical industrial monitoring

This study examines the application of drone-assisted infrared (IR) imaging with vision grayscale imaging and deep learning for enhanced abnormal detection in nuclear power plants. A scaled model, replicating the modern pressurized water reactor, facilitated the data collection for normal and abnormal conditions. A drone, equipped with dual vision and IR cameras, captured detailed operational imagery, crucial for detecting subtle anomalies within the plant's primary systems. Deep learning algorithms were deployed to interpret these images, aiming to identify component abnormals not easily discernible by traditional monitoring. The object detection model was trained to classify normal and abnormal component states within the facility, marked by color-coded bounding boxes for clarity. Models like YOLO and Mask R-CNN were evaluated for their precision in anomaly detection. Results indicated that the YOLO v8m model was particularly effective, showcasing high accuracy in both detecting and adapting to system anomalies, as validated by high mAP scores. The integration of drone technology with IR imaging and deep learning illustrates a significant stride toward automating abnormal detection in complex industrial environments, enhancing operational safety and efficiency. This approach has the potential to revolutionize real-time monitoring in safety–critical settings by providing a comprehensive, automated solution to abnormal detection.

SiCw strengthened WC-Y2O3 composite and its electrochemical corrosion in aggressive medium

The mechanical properties and corrosion resistance of binderless tungsten carbide largely determine its performance and lifespan in complex industrial environments. Herein, WC-1Y2O3-xSiCw (x = 0, 0.5, 1, and 2 wt%) was prepared using the hot-press sintering. The mechanical properties of the composite, as well as the corrosion resistance of optimized mechanical properties in an aggressive medium (0.05 M H2SO4 and 1 M NaOH) were investigated. Results showed that SiCw could significantly increase mechanical properties of the composite, but excessive content (2 wt%) leads to whiskers aggregation and lower properties. The flexural strength and fracture toughness are all increased to varying degrees with whisker content. WC-1Y2O3-0.5SiCw possesses the optimal mechanical properties. The Vickers hardness, fracture toughness, and flexural strength are 2331.24 HV30, 8.59 MPa·m1/2, and 1727 MPa, respectively. Electrochemical tests show that the corrosion resistance of the WC-1Y2O3-0.5SiCw in H2SO4 is higher than that in NaOH solution, and its corrosion resistance in H2SO4 solution increases, but that of the NaOH solution decreases with exposure time. Detailed characterization demonstrated that a dense WO3 film forms in the H2SO4 solution, resulting in a flatter corrosion surface. In contrast, the WC phase in NaOH solution dissolves, thus leading to the accelerated corrosion rate.

A Multi-Featured Factor Analysis and Dynamic Window Rectification Method for Remaining Useful Life Prognosis of Rolling Bearings.

Currently, the research on the predictions of remaining useful life (RUL) of rotating machinery mainly focuses on the process of health indicator (HI) construction and the determination of the first prediction time (FPT). In complex industrial environments, the influence of environmental factors such as noise may affect the accuracy of RUL predictions. Accurately estimating the remaining useful life of bearings plays a vital role in reducing costly unscheduled maintenance and increasing machine reliability. To overcome these problems, a health indicator construction and prediction method based on multi-featured factor analysis are proposed. Compared with the existing methods, the advantages of this method are the use of factor analysis, to mine hidden common factors from multiple features, and the construction of health indicators based on the maximization of variance contribution after rotation. A dynamic window rectification method is designed to reduce and weaken the stochastic fluctuations in the health indicators. The first prediction time was determined by the cumulative gradient change in the trajectory of the HI. A regression-based adaptive prediction model is used to learn the evolutionary trend of the HI and estimate the RUL of the bearings. The experimental results of two publicly available bearing datasets show the advantages of the method.

GBSG-YOLOv8n: A Model for Enhanced Personal Protective Equipment Detection in Industrial Environments

The timely and accurate detection of whether or not workers in an industrial environment are correctly wearing personal protective equipment (PPE) is paramount for worker safety. However, current PPE detection faces multiple inherent challenges, including complex backgrounds, varying target size ranges, and relatively low accuracy. In response to these challenges, this study presents a novel PPE safety detection model based on YOLOv8n, called GBSG-YOLOv8n. First, the global attention mechanism (GAM) is introduced to enhance the feature extraction capability of the backbone network. Second, the path aggregation network (PANet) structure is optimized in the Neck network, strengthening the model’s feature learning ability and achieving multi-scale feature fusion, further improving detection accuracy. Additionally, a new SimC2f structure has been designed to handle image features and more effectively improve detection efficiency. Finally, GhostConv is adopted to optimize the convolution operations, effectively reducing the model’s computational complexity. Experimental results demonstrate that, compared to the original YOLOv8n model, the proposed GBSG-YOLOv8n model in this study achieved a 3% improvement in the mean Average Precision (mAP), with a significant reduction in model complexity. This validates the model’s practicality in complex industrial environments, enabling a more effective detection of workers’ PPE usage and providing reliable protection for achieving worker safety. This study emphasizes the significant potential of computer vision technology in enhancing worker safety and provides a robust reference for future research regarding industrial safety.

Improved YOLOV8 Network and Application in Safety Helmet Detection

This paper proposes a research method to enhance the accuracy and real-time capability of helmet detection in complex industrial environments, aiming to address the engineering challenges of poor robustness and significant occurrences of both false positives and false negatives in existing detection methods. In this study, the C2F (faster version of CSP Bottleneck with two convolutions) module and FE (FasterNet with EMA) module are integrated into the network architecture of YOLOV8 to form a new attention mechanism module called C2F-FE. This module enhances the model’s perception of safety helmet targets by fusing feature information from different levels and incorporating attention mechanisms while reducing computational overhead. Furthermore, the model is trained and optimized on publicly available safety helmet datasets. Experimental results demonstrate that the improved model exhibits stronger robustness, achieving an accuracy rate of 94.6% and a mAP50 of 99.1% for safety helmet detection in complex construction scenarios, with an inference time of 0.7 ms.

Weld beads and defects automatic identification, localization, and size calculation based on a lightweight fully convolutional neural network

Weld quality inspection is a key aspect of modern manufacturing. Automatic identification, location, and dimensioning of welds and defects are necessary for complex industrial environments. Weld and defects dimensional data are important indicators for high precision assessment of weld quality, while precise positioning of welds and defects can provide important location data for subsequent processing. In this study, a novel weld and defects detection, location, and size calculation method is proposed based on deep fully convolutional neural networks. Firstly, 43 high-precision high-power laser weld plate scanned images are automatically cropped to 800 pixels × 800 pixels by the proposed sliding window sampling method for training, testing, and validation, while an image data enhancement method is proposed to quadruple the amount of data to improve the model performance. Next, the OU-net image semantic segmentation model is established to accurately identify different detection targets, which can achieve MIoU and MPA values of 78.15% and 85.18% on the validation set after sufficient training, respectively. Then, the proposed cropped segmentation image fusion recovery method is used to obtain the complete weld segmentation image for a complete display of the different inspection targets on the weld plate. Finally, a detection targets localization and size calculation method is applied to accurately obtain the position data and size data of different targets. The validation results show that the proposed detection strategy can locate and measure every weld seam, collapse, overlap, spatter, and hump on the weld plate.

An intelligent signal processing method against impulsive noise interference in AIoT

In complex industrial environments such as the Internet of Things in coal mines, large mechanical and electrical equipment can generate powerful impulsive noise, which can cause sudden errors. Because it is difficult to establish an accurate channel model, the performance of current error control techniques is limited. To enhance the reliability of information recovery in the Internet of Things in coal mines, the traditional method of shortening the communication distance between sensors is often utilized, but this can be costly. Therefore, this article proposes an intelligent signal processing method against impulsive noise interference that draws on the concept of the Artificial Intelligence of Things (AIoT) and incorporates deep learning technology. This method replaces the traditional sensor signal processing module with a Convolutional Neural Network (CNN), which learns the intricate mapping relationship between transmitted information and sensor signals in impulsive noise environments. Simulation results demonstrate that the proposed method outperforms the traditional sensor signal processing method in three impulsive noise environments by achieving a lower Bit Error Rate (BER). Moreover, this method adopts an improved lightweight neural network, which is more conducive to the deployment of mobile terminals in the Internet of Things.

Q-learning based multi-objective immune algorithm for fuzzy flexible job shop scheduling problem considering dynamic disruptions

Confronted with complex industrial environments, dynamic disruptions like new job arrival and machine breakdown bring significant challenges to the robustness and stability of the manufacturing process, making the static production depart from the original scheduling scheme. To address this problem, a flexible job shop scheduling problem with fuzzy processing time, dynamic disruptions, and variable processing speeds is considered simultaneously. As well as three objectives of maximum completion time, total energy consumption, and average agreement index are demonstrated in this study. Then, a predictive-reactive dynamic/static rescheduling model is developed, where the off-line based mixed integer linear programming model and the on-line based rescheduling heuristics are proposed. Next, a multi-objective immune algorithm combined with a Q-learning algorithm (Q-MOIA) is developed. In the proposed algorithm, an active decoding heuristic based on the interval insertion mechanism is used to optimize the initial solutions. After that, the clonal selection-based immune algorithm and the Q-learning algorithm are adopted to improve the exploration and exploitation capabilities, respectively, where four objective-driven neighborhood structures are designed. Eventually, extensive computational experiments were conducted on 27 instances under static and dynamic scenarios to demonstrate the superiority and stability of the proposed predictive-reactive dynamic/static rescheduling model and the Q-MOIA. Comparative analysis with four state-of-the-art approaches revealed that proposed Q-MOIA outperformed in approximately 51.9, 66.7, and 83.3 % of the instances for the three multi-objective metrics.

Augmented reality and indoor positioning based mobile production monitoring system to support workers with human-in-the-loop

With the increasing demand for product customization, the exponential increase in large-scale and small-batch production orders has yielded new challenges for the capacity of traditional production lines. Due to the information redundancy and complex production process on the production site of the factory, the traditional monitoring system is incapable of meeting the information interaction requirements between the factory management level and the executive level. Meanwhile, the traditional augmented reality (AR) based on image recognition is not suitable in the complex industrial environments. To address the gap, we propose a novel mobile production monitoring system (AIMPMs) with human-in-the-loop control by leveraging the cutting-edge AR and indoor positioning technique. In the proposed system, Ultra-wideband (UWB) and Inertial Measurement Unit (IMU) fusion indoor positioning technology is proposed, which provides accurate indoor positioning information for the production factors in the factory. Subsequently, we build the lightweight indoor map for positioning that can serve as the location reference and path planning, and a nearest-neighbor decision algorithm with double rejection decision (NN-DRD) is proposed to match the positioning features to trigger virtual monitoring information. Finally, the AIMPMs is applied in a hydraulic cylinder factory to verify its enforceability and effectiveness, and the human factors evaluation model and index system are constructed to evaluate two systems, (1) the mobile monitoring system based on positioning information, and (2) the mobile monitoring system based on image recognition, respectively. The experimental results indicate that after the application of AIMPMs, the physiological and mental fatigue of the production personnel is immensely decreased. Therefore, the system realizes a smarter and highly humanized human-machine interaction mode with human-in-the-loop control.

Reinforcement Learning-Based Tracking Control for Two Time-Scale Looper Hydraulic Servo Systems

Abstract Looper hydraulic servosystems (HSSs) are used in a variety of complex industrial environments. It is worth noting that there exists two-time-scale phenomenon in looper hydraulic servosystems. In this technical brief, the model of the looper hydraulic servosystems with two-time-scale phenomenon is established. In the framework of reinforcement learning, an optimal tracking control policy for looper hydraulic servosystems is proposed, which considers the effects of two-time-scale phenomena on the system. Then, by analyzing the change of looper angle in the actual industrial production process, the designed controller has a faster transition time compared to the proportional integral differential controller. In addition, the convergence of the optimal control solution obtained by using the proposed algorithm is proved. Finally, the effectiveness of the proposed method is verified by the simulation.

The artificial intelligence technologies in Industry 4.0: A taxonomy, approaches, and future directions

Industry 4.0 transforms the manufacturing sector with dynamic, networked, complex industrial environments. These environments generate vast amounts of data and require technology and Artificial Intelligence (AI) to achieve intelligent, efficient, and sustainable production processes. This paper comprehensively reviews 45 articles on AI and Industry 4.0 integration. We propose a taxonomy for AI in Industry 4.0 and classify approaches into Industry 4.0 design and product quality control methods. Our analysis shows that 58% of papers use product quality control methods. Besides, this paper identifies challenges and open issues by illuminating the current landscape. The findings showed that machine learning is the most common AI method for improving Industry 4.0 with 41%, and Python is the most used tool for simulation.