Industrial Tasks Research Articles

ContextMix: A context-aware data augmentation method for industrial visual inspection systems

While deep neural networks have achieved remarkable performance, data augmentation has emerged as a crucial strategy to mitigate overfitting and enhance network performance. These techniques hold particular significance in industrial manufacturing contexts. Recently, image mixing-based methods have been introduced, exhibiting improved performance on public benchmark datasets. However, their application to industrial tasks remains challenging. The manufacturing environment generates massive amounts of unlabeled data on a daily basis, with only a few instances of abnormal data occurrences. This leads to severe data imbalance. Thus, creating well-balanced datasets is not straightforward due to the high costs associated with labeling. Nonetheless, this is a crucial step for enhancing productivity. For this reason, we introduce ContextMix, a method tailored for industrial applications and benchmark datasets. ContextMix generates novel data by resizing entire images and integrating them into other images within the batch. This approach enables our method to learn discriminative features based on varying sizes from resized images and train informative secondary features for object recognition using occluded images. With the minimal additional computation cost of image resizing, ContextMix enhances performance compared to existing augmentation techniques. We evaluate its effectiveness across classification, detection, and segmentation tasks using various network architectures on public benchmark datasets. Our proposed method demonstrates improved results across a range of robustness tasks. Its efficacy in real industrial environments is particularly noteworthy, as demonstrated using the passive component dataset.

Applying Density-Based Clustering for the Analysis of Emission Events in Real Driving Emissions Calibration

Further reducing greenhouse gas and pollutant emissions from road vehicles is a major task for the automotive industry. Stricter regulations regarding emissions and fleet fuel consumption require the continuous development of new powertrains and methods. In particular, the combination of hybrid powertrains on the technical side and the focus on real driving emissions (RDE) on the legislative side pose significant challenges to the vehicle calibration process. Against this background, new test methods and environments are being investigated to counteract the high number of interactions between hybrid drive systems and quasi-infinite test conditions due to RDE. Complementary to new test environments, innovative methods for data analysis are needed that allow the exploitation of the complete potential of measurement data. The application of such a method in the field of emission calibration is presented in this paper. For this purpose, a clustering method (HDBSCAN) is applied to critical sequences from emission tests. Within this presentation, the clustering process is based on a single signal only. This paper shows how signals of various characteristics can be processed with dynamic time warping and generically structured with the clustering method used. Here, 959 single events are automatically categorized into 24 clusters. This provides a new basis for system evaluation, enabling the automatic identification, categorization, and prioritization of calibration weaknesses. Using twelve signals of different characteristics, the generic usability of the clustering method is demonstrated.

Combining Synthetic Images and Deep Active Learning: Data-Efficient Training of an Industrial Object Detection Model.

Generating synthetic data is a promising solution to the challenge of limited training data for industrial deep learning applications. However, training on synthetic data and testing on real-world data creates a sim-to-real domain gap. Research has shown that the combination of synthetic and real images leads to better results than those that are generated using only one source of data. In this work, the generation of synthetic training images via physics-based rendering is combined with deep active learning for an industrial object detection task to iteratively improve model performance over time. Our experimental results show that synthetic images improve model performance, especially at the beginning of the model's life cycle with limited training data. Furthermore, our implemented hybrid query strategy selects diverse and informative new training images in each active learning cycle, which outperforms random sampling. In conclusion, this work presents a workflow to train and iteratively improve object detection models with a small number of real-world images, leading to data-efficient and cost-effective computer vision models.

A Laboratory-Based Multidisciplinary Approach for Effective Education and Training in Industrial Collaborative Robotics

The rapid evolution of robotics across various sectors, including healthcare, manufacturing, and domestic applications, has underscored a significant workforce skills gap. The shortage of qualified professionals in the labor market has had adverse effects on production capacities. Therefore, the significance of education and training for cultivating a skilled workforce cannot be overstated. This research work presents the development of a pedagogical approach centered on laboratory infrastructure designed specifically with multidisciplinary technologies and strategic human–machine interaction protocols to enhance learning in industrial robotics courses. Progressive competencies in laboratory protocols are developed, focusing on programming and simulating real-world industrial robotics tasks, to bridge the gap between theoretical education and practical industrial applications for higher education students. The proposed infrastructure includes a user-configurable maze comprising different colored elements, defining starting points, endpoints, obstacles, and varying track sections. These elements foster a dynamic and unpredictable learning environment. The infrastructure is fabricated using Computer Numerical Control (CNC) machining and 3D printing techniques. A collaborative robot, the Universal Robots UR3e, is used to navigate the maze and solve the track with advanced computer vision and human–machine communication. The amalgamation of practical experience and collaborative robotics furnishes students with hands-on experience, equipping them with the requisite skills for effective programming and manipulation of robotic devices. Empowering human–machine interaction and human–robot collaboration assists in addressing the industry’s demand for skilled labor in operating collaborative robotic manipulators.

Flow in human-robot collaboration-multimodal analysis and perceived challenge detection in industrial scenarios.

Introduction: Flow state, the optimal experience resulting from the equilibrium between perceived challenge and skill level, has been extensively studied in various domains. However, its occurrence in industrial settings has remained relatively unexplored. Notably, the literature predominantly focuses on Flow within mentally demanding tasks, which differ significantly from industrial tasks. Consequently, our understanding of emotional and physiological responses to varying challenge levels, specifically in the context of industry-like tasks, remains limited. Methods: To bridge this gap, we investigate how facial emotion estimation (valence, arousal) and Heart Rate Variability (HRV) features vary with the perceived challenge levels during industrial assembly tasks. Our study involves an assembly scenario that simulates an industrial human-robot collaboration task with three distinct challenge levels. As part of our study, we collected video, electrocardiogram (ECG), and NASA-TLX questionnaire data from 37 participants. Results: Our results demonstrate a significant difference in mean arousal and heart rate between the low-challenge (Boredom) condition and the other conditions. We also found a noticeable trend-level difference in mean heart rate between the adaptive (Flow) and high-challenge (Anxiety) conditions. Similar differences were also observed in a few other temporal HRV features like Mean NN and Triangular index. Considering the characteristics of typical industrial assembly tasks, we aim to facilitate Flow by detecting and balancing the perceived challenge levels. Leveraging our analysis results, we developed an HRV-based machine learning model for discerning perceived challenge levels, distinguishing between low and higher-challenge conditions. Discussion: This work deepens our understanding of emotional and physiological responses to perceived challenge levels in industrial contexts and provides valuable insights for the design of adaptive work environments.

A novel AR recoloring technique to enhance operator performance on inspection tasks in Industry 4.0 environments.

Over the past few years, the manufacturing industry has increasingly embraced Augmented Reality (AR) for inspecting real products, yet faces challenges in visualization modalities. In fact, AR content presentation significantly impacts user performance, especially when virtual object colors lack real-world context. Additionally, the lack of studies in this area compounds uncertainty about visualization effects on user performance in inspection tasks. This study introduces a novel AR recoloring technique to enhance user performance during industrial assembly inspection tasks. This technique automatically recolors virtual components based on their physical counterparts, improving distinctiveness. Experimental comparisons with AR experts and representative users, using objective and subjective metrics, demonstrate the proposed AR recoloring technique enhances task performance and reduces mental burden during inspection activities. This innovative approach outperforms established methods like CAD and random modes, showcasing its potential for advancing AR applications in manufacturing, particularly in the inspection of products.

Neuromuscular Interfacing for Advancing Kinesthetic and Teleoperated Programming by Demonstration of Collaborative Robots.

This study addresses the challenges of Programming by Demonstration (PbD) in the context of collaborative robots, focusing on the need to provide additional degrees of programming without hindering the user's ability to demonstrate trajectories. PbD enables an intuitive programming of robots through demonstrations, allowing non-expert users to teach robot skills without coding. The two main PbD modalities, observational and kinesthetic, have limitations when it comes to programming the diverse functionalities offered by modern collaborative robots. To overcome these limitations, the study proposes the use of a wearable human-robot interface based on surface Electromyography (sEMG) to measure the forearm's muscle co-contraction level, enabling additional programming inputs through hand stiffening level modulations without interfering with voluntary movements. Vibrotactile feedback enhances the operator's understanding of the additional programming inputs during PbD tasks. The proposed approach is demonstrated through experiments involving a collaborative robot performing an industrial wiring task. The results showcase the effectiveness and intuitiveness of the interface, allowing simultaneous programming of robot compliance and gripper grasping. The framework, applicable to both teleoperation and kinesthetic teaching, demonstrated effectively in an industrial wiring task with a 100% success rate over the group of subjects. Furthermore, the presence of vibortactile feedback showed an average decrease of programming errors of 33%, and statistical analyses confirmed the subjects' ability to correctly modulate co-contraction levels. This innovative framework augments programming by demonstration by integrating neuromuscular interfacing and introducing structured programming logics, providing an intuitive human-robot interaction for programming both gripper and compliance in teleoperation and kinesthetic teaching.

A Multi-Criteria Group Decision-Making Approach For Robot Selection Using Interval-Valued Intuitionistic Fuzzy Information And Aczel-Alsina Bonferroni Means

The process of identifying the most appropriate robot for a particular industrial task has grown challenging and more difficult in the fast-paced environment. It is merely driven by the complex evolution and continuous integration of modern characteristics and advanced features by various suppliers. Industrial robots are now widely available in the marketplace, each possessing a distinctive collection of skills, attributes, and requirements. However, the selection of optimal robots is heavily influenced by factors such as the manufacturing environment, product design, production system, and overall cost considerations. These factors directly impact the decision-making process. The ultimate goal for the decision maker is to pinpoint and choose the most suitable robot, capable of delivering the desired output while minimizing costs and catering to the specific requirements of the industry. So, to consider it, in this paper, the hybrid structure of the Aczel-Alsina (AA) and Bonferroni mean (BM) operators for interval-valued intuitionistic fuzzy (IVIF) environment has been proposed, which can show the interrelationship between the multiple criteria and assist the expertise in decision-making (DM) process. Moreover, the algorithm and methodology for the multi-criteria group decision-making (MCGDM) problem have been defined which is further utilized by solving a real-world problem to demonstrate the effectiveness and validity of the proposed method. At last, the comparison analysis between prior and proposed studies has been displayed, and then followed by the conclusion of the result.

Impact of a passive upper-body exoskeleton on muscular activity and precision in overhead single and dual tasks: an explorative randomized crossover study.

Tasks performed at or above head height in industrial workplaces pose a significant challenge due to their association with musculoskeletal disorders. Upper-body exoskeletons have been identified as a potential solution for mitigating musculoskeletal loads and fighting against excessive muscular fatigue. However, the influence of such support on fine motor control, as well as on cognitive-motor interference, has received limited attention thus far. Therefore, this crossover randomized study aimed to investigate the impact of the use of a passive upper-body exoskeleton in the presence of muscular fatigue or not. Additionally, focusing on differences between single (ST) and dual (DT) industrial tasks consisting of overhead speed and accuracy exercises. In both scenarios, N = 10 participants (5 male/5 female) engaged in an overhead precision task using a nail gun to precisely target specific areas on three differently sized regions, based on Fitts' law paradigm (speed-accuracy trade-off task). This was done with and without the passive upper-body exoskeleton, before and immediately after a fatiguing exercise of shoulder and leg muscles. In addition, a second task (dual-task, DT) was carried out in which the occurrence of an auditory signal had to be counted. The main outcomes were muscular activation of the shoulder girdle as well as the time to perform speed-accuracy tasks of different difficulty indexes (calculated by means of Fitts' law). In the absence of fatigue, the exoskeleton did not affect the speed-accuracy trade-off management of participants in the single task, but it did in the dual-task conditions. However, after muscle fatigue, the speed-accuracy trade-off was differently affected when comparing its execution with or without the exoskeleton. In general, the dual task resulted in longer times to perform the different tasks, whether it was with or without the exoskeleton. Furthermore, the use of the exoskeleton decreased muscle activity, which is associated with less physical effort, but only significantly for the M. deltoideus and M. trapezius when compared by tasks. Overall, these study findings highlight the potential supportive effects of using an upper-body exoskeleton for industrial overhead tasks.

A multi-facet approach to functional and ergonomic assessment of passive exoskeletons

Seeing the innovative character of exoskeletal devices, it appears that there is yet progress to be made in adopting exoskeletons to industry, and to the different users’ characteristics and needs. Their validation is mainly focused on technological aspects and seldom researchers presented methodological approaches including both physical and cognitive impact on operators, synergically with functional issues. This study proposes a multi-faced approach to investigate the impact of passive exoskeletons on functional and ergonomic aspects providing a preliminary experimental validation with an industrial task. The approach assesses the physical comfort, muscular fatigue, cognitive workload, and users’ technological acceptance, to map the overall user experience (UX). Participants were monitored by a mechanomyogram (MMG) and a motion capture (MoCap) system and a UX questionnaire was administered. Results demonstrated the effectiveness of passive exoskeletons regarding the specific task being strictly related to a correct definition of the level of assistance for each user.

In-situ ergonomics assessment of industrial manual tasks through AR-headset integrated sensor-systems

This paper presents an approach that is based on the use of sensors integrated into Augmented Reality (AR) headsets to acquire data for in-situ ergonomics assessment of industrial tasks. In the context of human factors evaluation, it is typically required to collect data related to anthropometric characteristics of the operators e.g. height, length of arms, width of shoulders etc., as well as data related to the execution of manufacturing tasks including working postures, task repetition, task duration, and so forth. To this end, it is needed to develop a plug and play system to collect ergonomics data from the field. The proposed approach relies on head-mounted AR devices and their integrated RGB and Depth sensors, image and point cloud processing, and skeleton tracking techniques to calculate the required data, which are later visualized in a user-friendly interface. The proposed approach was tested in a case study where the objective was to identify the proper mechanical adjustments of an exoskeleton on the spot for comfortable fitting for each operator. Five operators participated in the testing that was held under varying lighting conditions. The solution showcased improved efficiency in decision-making compared to the commonplace procedures for skeleton fitting, faster setup time compared to typical camera-based approaches, and satisfactory performance.

Use of augmented reality for iterative robot program optimisation in robot-automated series production processes

The inspection, analysis, and modification of robot programs in series production processes is a challenging recurring task in the automotive industry. Especially the holistic analysis of detected defects and the subsequent derivation and application of corrective measures to the underlying robot program is a complex task requiring extensive knowledge in multiple fields. Existing scientific literature has demonstrated that the utilisation of Augmented Reality (AR) assisted robot programming systems (ARRPS) improves programming efficiency as well as intuitiveness when compared to traditional programming methods. However, the accuracy especially of mobile AR devices limits the applicability of AR in industrial applications. We propose to leverage the iterative nature of the optimisation process common in series production and utilise the visual capabilities of the human worker to define corrective measures based on the inspected real process result. This mitigates the impact of limited tracking accuracy as the optimisation is based on the visual perception of the user and thus decoupled from the accuracy of the employed AR device. We implemented our system based on the requirements of a sealing workstation of the Mercedes-Benz Group in Germany. A conducted experiment validates the functionality and indicates that the presented strategy of iterative AR assisted robot programming enables optimisations with a similar accuracy as the Teach-In programming as well as a significantly reduced programming time. This suggests that utilising an ARRPS can be highly beneficial in the inspection, analysis, and modification of sealing applications common in automotive series production processes. As the presented iterative AR assisted robot programming is neither limited to sealing nor to the automotive sector it can be adapted for various robot-automated applications like path-welding, gluing, or painting and can extend its utility to related sectors such as aviation or the marine industry.

A refined decision-making method for orderly power consumption in virtual power plants considering load characteristics

Abstract In recent years, electricity consumption has been increasing, while the supply of electricity is relatively lagging, and power shortages often occur. Therefore, the management of organized power consumption is an urgent task for the current power industry. Based on the characteristics of power load, this paper constructs a decision-making management framework for orderly power consumption by considering four aspects: generation, grid side, demand side, and social benefit. Through the various types of flexible load resources aggregated within the virtual power plant, it participates in regulating its power, effectively reduces the peak-valley difference, and realizes refined, orderly power consumption management. Finally, based on the multi-objective optimization model, refined decision-making for orderly power consumption with the participation of virtual power plants is constructed and compared to traditional decision-making for orderly power consumption. The results show that compared with conventional decision-making, the peak-to-valley difference is reduced by 8.7789 WM, the economic benefit of the power enterprise is improved by 228,300 yuan, and customer satisfaction is enhanced by 0.1613, which verifies the effectiveness of the decision-making method of the sequential power consumption in this paper.

Влияние алюмосиликатных ценосфер на структуру и свойства эластомерных композиционных материалов на основе этилен-пропилен-диеновых эластомеров

The use of industrial waste to develop new value-added materials and reduce their environmental impact is one of the most important tasks of science and industry, which can largely solve the problem of waste pollution. This work is aimed at studying the effect of different concentrations of fly ash aluminosilicate cenospheres on the structure and properties of elastomeric composites. In this work, using laboratory rollers, composite materials based on ethylene-propylene-diene rubber (EPDM-40) with different mass fractions of fly ash (10, 20 and 30 %) were obtained. Using the method of optical microscopy, the microstructure of mixtures of EPDM and aluminosilicate cenospheres was studied and it was shown that the filler content of more than 30% increases the content of larger cenosphere agglomerates in the structure, which indicates interfacial separation in mixtures, probably due to the fact that mechanical mixing on mixing equipment does not allow to achieve a uniform distribution of the filler throughout the elastomeric matrix. The IR spectra show the appearance of new absorption bands in the region. 1400 – 800 cm–1, corresponding to Si – O – Si stretching vibrations present in aluminosilicate cenospheres. According to the thermogravimetry data of the studied compositions, the introduction of aluminosilicate cenospheres contributed to a slight increase in the thermal stability of the tested composition with a cenosphere content of more than 30%. The influence of the concentration of aluminosilicate cenospheres on the resistance of composites to aggressive media was analyzed and it was found that the introduction of a cenosphere filler in an amount of 10 to 30 % in a mixture based on EPDM can increase the oil and petrol resistance of materials.

HSDNet: a poultry farming model based on few-shot semantic segmentation addressing non-smooth and unbalanced convergence.

Poultry farming is an indispensable part of global agriculture, playing a crucial role in food safety and economic development. Managing and preventing diseases is a vital task in the poultry industry, where semantic segmentation technology can significantly enhance the efficiency of traditional manual monitoring methods. Furthermore, traditional semantic segmentation has achieved excellent results on extensively manually annotated datasets, facilitating real-time monitoring of poultry. Nonetheless, the model encounters limitations when exposed to new environments, diverse breeding varieties, or varying growth stages within the same species, necessitating extensive data retraining. Overreliance on large datasets results in higher costs for manual annotations and deployment delays, thus hindering practical applicability. To address this issue, our study introduces HSDNet, an innovative semantic segmentation model based on few-shot learning, for monitoring poultry farms. The HSDNet model adeptly adjusts to new settings or species with a single image input while maintaining substantial accuracy. In the specific context of poultry breeding, characterized by small congregating animals and the inherent complexities of agricultural environments, issues of non-smooth losses arise, potentially compromising accuracy. HSDNet incorporates a Sharpness-Aware Minimization (SAM) strategy to counteract these challenges. Furthermore, by considering the effects of imbalanced loss on convergence, HSDNet mitigates the overfitting issue induced by few-shot learning. Empirical findings underscore HSDNet's proficiency in poultry breeding settings, exhibiting a significant 72.89% semantic segmentation accuracy on single images, which is higher than SOTA's 68.85%.

Роль закупочной деятельности в уменьшении углеродного следа в Арктической зоне Российской Федерации

One of the main tasks of the modern energy industry is not only to maximize the efficient use of natural energy resources and preserve their diversity, but also to increase the potential of this sector in the stable growth of the state's economy, improving the quality of life of the population and maintaining a favorable environmental situation in the country. Decarbonization, which implies a reduction in carbon dioxide (CO<sub>2</sub>) and methane emissions, is a key challenge facing the global energy industry. The purpose of this study is to assess greenhouse gas emissions produced as a result of the activities of the largest energy companies operating in the Arctic zone of the Russian Federation, namely PJSC LUKOIL and PJSC NOVATEK through the analysis of their policies and measures to reduce the carbon footprint, preserve biological diversity and develop the northern region of Russia, identify links between the conduct of procurement activities and support for the sustainable development of the enterprise, including in the environmental field. In the course of the work, a comparative analysis of the activities of companies in the Arctic zone was carried out, their current and prospective projects were described, data on greenhouse gas emissions of the enterprises in question were structured by year, their procurement policy, the necessary requirements for potential suppliers, implemented educational programs and activities for the development of professional skills of personnel were considered, and additional tools were proposed that they will be able to contribute not only to improving the sustainable development of companies in the Arctic, but also, as a result, conservation of natural diversity in the northern part of the Russian Federation.

Automatized, holistic use of information from milling processes

The improvement of the various CAx (information) process chains along the product lifecycle is a crucial task for the production industry to stay competitive. The collection, processing and visualization of information from different sources (process-, tools and quality) along the CAD/CAM process chain for different stakeholders (designer, programmer, process experts and others) is the main objective of the research project AUTODAT - AUTOmatized, holistic use of production DATa, executed by Premium AEROTEC and Jade University as a cooperative project partners.This contribution illustrates the achievements of a joined approach of the three disciplines: (1) condition monitoring, based on vibration analysis, (2) information modelling as a conceptional basis for the development and (3) visualization of quality and process data using user experience design methods. The architecture of the information flow will be presented as a feedback loop to support early stages of product lifecycle – design and NC programming and simulation. The industrial challenges and company-specific best practices for the implementation are shown as well. Through the joint effort of disciplines, a gap in the information flow could be identified and closed. The flow of process, tool and quality information was established. Thereby, the whole process was improved.

Study of the influence of corona discharge on the efficiency of oxidation of organic impurities using ozone

Purification of air containing volatile organic compounds is an important task in many industries. This article examines the effect of corona discharge on the efficiency of the oxidation of organic mixtures. As a result of our research, we found that the consumption of ozone for the oxidation of various components of the mixture is approximately the same. This means that other particles, including water vapor, also take part in oxidation reactions.

Real-time processing and optimization strategies for IoT data streams

Abstract With the development of industrial IoT and the arrival of smart manufacturing, the field of edge computing has gained more and more attention. However, traditional industrial computing scenarios relying on industrial clouds make data latency a greater challenge. In this paper, for the contradiction between edge devices and task resource allocation encountered in edge computing scenarios in smart manufacturing, we propose an industrial internet task scheduling model for smart manufacturing and introduce a scheduling node state matrix to realize the state management of each scheduling subtask. Aiming at the problem of multiple tasks seizing resources in a complex, intelligent manufacturing environment, the study combines the caching mechanism to realize the task offloading computational processing of order scheduling, in which the caching mechanism is used to solve the problem of computational resource limitations at the edge. It is found through simulation that when the computational task factor ξk =2 is larger, more offloading power is allowed to be transmitted to the edge ser ver for computation. For computational tasks with smaller task factor ξk , the device tends to allocate more computational rate to that computational task. Eventually the data queue length will be continuously reduced and the data queue is concentrated in the interval of very small values, this result verifies that the task scheduling algorithm is able to perform task scheduling efficiently and reduce the latency.

Industrial design change task allocation method for complex product manufacturing based on implementation intention matching

Industrial design change task allocation method for complex product manufacturing based on implementation intention matching