Refine
Document Type
- Conference proceeding (847) (remove)
Language
- English (847) (remove)
Is part of the Bibliography
- yes (847)
Institute
- Informatik (459)
- Technik (220)
- ESB Business School (137)
- Texoversum (21)
- Life Sciences (11)
- Zentrale Einrichtungen (2)
Publisher
- IEEE (222)
- Springer (136)
- Gesellschaft für Informatik (42)
- ACM (38)
- Hochschule Reutlingen (31)
- Association for Information Systems (AIS) (29)
- IARIA (19)
- SSRN (18)
- SCITEPRESS (14)
- VDE Verlag GmbH (14)
Human pose estimation (HPE) is integral to scene understanding in numerous safety-critical domains involving human-machine interaction, such as autonomous driving or semi-automated work environments. Avoiding costly mistakes is synonymous with anticipating failure in model predictions, which necessitates meta-judgments on the accuracy of the applied models. Here, we propose a straightforward human pose regression framework to examine the behavior of two established methods for simultaneous aleatoric and epistemic uncertainty estimation: maximum a-posteriori (MAP) estimation with Monte-Carlo variational inference and deep evidential regression (DER). First, we evaluate both approaches on the quality of their predicted variances and whether these truly capture the expected model error. The initial assessment indicates that both methods exhibit the overconfidence issue common in deep probabilistic models. This observation motivates our implementation of an additional recalibration step to extract reliable confidence intervals. We then take a closer look at deep evidential regression, which, to our knowledge, is applied comprehensively for the first time to the HPE problem. Experimental results indicate that DER behaves as expected in challenging and adverse conditions commonly occurring in HPE and that the predicted uncertainties match their purported aleatoric and epistemic sources. Notably, DER achieves smooth uncertainty estimates without the need for a costly sampling step, making it an attractive candidate for uncertainty estimation on resource-limited platforms.
In recent years, the demand for accurate and efficient 3D body scanning technologies has increased, driven by the growing interest in personalised textile development and health care. This position paper presents the implementation of a novel 3D body scanner that integrates multiple RGB cameras and image stitching techniques to generate detailed point clouds and 3D mesh models. Our system significantly enhances the scanning process, achieving higher resolution and fidelity while reducing the cost, time and effort required for data acquisition and processing. Furthermore, we evaluate the potential use cases and applications of our 3D body scanner, focusing on the textile technology and health sectors. In textile development, the 3D scanner contributes to bespoke clothing production, allowing designers to construct made-to-measure garments, thus minimising waste and enhancing customer satisfaction through fitting clothing. In mental health care, the 3D body scanner can be employed as a tool for body image analysis, providing valuable insights into the psychological and emotional aspects of self-perception. By exploring the synergy between the 3D body scanner and these fields, we aim to foster interdisciplinary collaborations that drive advancements in personalisation, sustainability, and well-being.
Patterns are virtually simulated in 3D CAD programs before production to check the fit. However, achieving lifelike representations of human avatars, especially regarding soft tissue dynamics, remains challenging. This is mainly since conventional avatars in garment CAD programs are simulated with a continuous hard surface and not corresponding to the human physical and mechanical body properties of soft tissue. In the real world, the human body’s natural shape is affected by the contact pressure of tight-fitting textiles. To verify the fit of a simulated garment, the interactions between the individual body shape and the garment must be considered. This paper introduces an innovative approach to digitising the softness of human tissue using 4D scanning technology. The primary objective of this research is to explore the interactions between tissue softness and different compression levels of apparel, exerting pressure on the tissue to capture the changes in the natural shape. Therefore, to generate data and model an avatar with soft body physics, it is essential to capture the deform ability and elasticity of the soft tissue and map it into the modification options for a simulation. To aim this, various methods from different fields were researched and compared to evaluate 4D scanning as the most suitable method for capturing tissue deformability in vivo. In particular, it should be considered that the human body has different deformation capabilities depending on age, the amount of muscle and body fat. In addition, different tissue zones have different mechanical properties, so it is essential to identify and classify them to back up these properties for the simulation. It has been shown that by digitising the obtained data of the different defined applied pressure levels, a prediction of the deformation of the tissue of the exact person becomes possible. As technology advances and data sets grow, this approach has the potential to reshape how we verify fit digitally with soft avatars and leverage their realistic soft tissue properties for various practical purposes.
Analog integrated circuit sizing still relies heavily on human expert knowledge as previous automation approaches have not found wide-spread acceptance in industry. One strand, the optimization-based automation, is often discarded due to inflated constraining setups, infeasible results or excessive run times. To address these deficits, this work proposes a alternative optimization flow featuring a designer’s intuition for feasible design spaces through integration of expert knowledge based on the gm/ID-method. Moreover, the extensive run times of simulation-based optimization flows are overcome by incorporating computationally efficient machine learning methods. Neural network surrogate models predicting eleven performance parameters increase the evaluation speed by 3 400× on average compared to a simulator. Additionally, they enable the use of optimization algorithms dependent on automatic differentiation, that would otherwise be unavailable in this field. First, an up to 4× more efficient way for sampling training data based on the aforementioned space is detailed. After presenting the architecture and training effort regarding the surrogate models, they are employed as part of the objective function for sizing three operational amplifiers with three different optimization algorithms. Additionally, the benefits of using the gm/ID-method become evident when considering technology migration, as previously found solutions may be reused for other technologies.
Facing ever-looming climate change, studying the drivers for individuals' Information Systems (IS) Use to reduce environmental harm gains momentum. While extant research on the antecedents of sustainable IS Use has focused on specific theories, interventions, contexts, and technologies, a holistic understanding has become increasingly elusive, with a synthesis remaining absent. We employ a systematic literature review methodology to shed light on the driving antecedents for sustainable IS Use among individual consumers. Our results build on findings of 29 empirical studies drawn from 598 articles retrieved from our premier outlets and a forward/backward search. The analysis reveals six salient complementary antecedents: Relief, Empowerment, Default, User-centricity, Salience, and Encouragement. We recommend considering these concepts when developing, deploying, promoting, or regulating digital technologies to mitigate individual consumers' emissions. Along with memorable and implementable concepts, our theoretical framework offers a novel conceptualization and four promising avenues for researchers on sustainable IS Use.
This article presents a modified method of performing power flow calculations as an alternative to pure energy-based simulations of off-grid hybrid systems. The enhancement consists in transforming the scenario-based power flow method into a discrete time-dependent algorithm with the inclusion of bus and controller dynamics.
Smart cities are considered data factories that generate an enormous amount of data from various sources. In fact data is the backbone of any smart services. Therefore, the strategic beneficial handling of this digital capital is crucial for cities. Some smart city pioneers have already written down their approach to data in the form of data strategies, but what should a city's data strategy include, and how can the goals and measures defined in the strategies be operationalized? This paper addresses these questions by looking closely at the data strategies of cities in Germany and the top three countries in the EU Digital Economy and Society Index. The in-depth analysis of 8 city data strategies has yielded 11 dimensions that cities should consider in their data strategy. These are relevance of data, principles, methods, data sharing, technology, data culture, data ethics, organizational structure, data security and privacy, collaborations, data literacy. In addition, data governance is a concept to put these 11 strategic dimensions into practice through standardization measures, training programs, and defining roles and responsibilities by developing a data catalog.
Platforms feature increasingly complex architectures with regard to interconnecting with other digital platforms as well as with a variety of devices and services. This development also impacts the structure of digital platform ecosystems and forces providers of these services, devices, and services to incorporate this complexity in their decision-making. To contribute to the existing body of knowledge on measuring ecosystem complexity, the present research proposes two key artefacts based on ecosystem intelligence: On the one hand, complementarity graphs represent ecosystems with an ecosystem's functional modules as vertices and complementarities as edges. The nodes carry information about the category membership of the module. On the other hand, a process is suggested that can collect important information for ecosystem intelligence using proxies and web scraping. Our approach allows replacing data, which today is largely unavailable due to competitive reasons. We demonstrated the use of the artefacts in category-oriented complementarity maps that aggregate the information from complementarity graphs and support decision-making. They show which combination of module categories creates strong and weak complementarities. The paper evaluates complementarity maps and the data collection process by creating category-oriented complementarity graphs on the Alexa skill ecosystem and concludes with a call to pursue more research based on functional ecosystem intelligence.
The proliferation of smart technologies transforms the way individual consumers perform tasks. Considerable research alludes that smart technologies are often related to domestic energy consumption. However, it remains unclear how such technologies transform tasks and thereby impact our planet. We explore the role of technological smartness in personal day-to-day tasks that help create a more sustainable future. In the absence of theory, but facing extensive changes in everyday life enabled by smart technologies, we draw on phenomenon-based theorizing (PBT) guidelines. As anchor, we refer to task endogeneity related to task-technology fit theory (TTF). As infusion, we employ theory on public goods. Our model proposes novel relations between the concepts of smart autonomy and -transparency with sustainable task outcomes, mediated by task convenience and task significance. We discuss some implications, limitations, and future research opportunities.
Smart factories, driven by the integration of automation and digital technologies, have revolutionized industrial production by enhancing efficiency, productivity, and flexibility. However, the optimization and continuous improvement of these complex systems present numerous challenges, especially when real-world data collection is time-consuming, expensive, or limited. In this paper, we propose a novel method for semi-automated improvement of smart factories using synthetic data and cause-effect-relations, while incorporating the aspect of self-organization. The method leverages the power of synthetic data generation techniques to create representative datasets that mimic the behaviour of real-world manufacturing systems. These synthetic datasets serve together with the cause-and-effect relationships as a valuable resource for factory optimization, as they enable extensive experimentation and analysis without the constraints of limited or costly real-world data. Furthermore, the method embraces the concept of self organization within smart factories. By allowing the system to adapt and optimize itself based on feedback from the synthetic data, cause-effect-relationships, the factory can dynamically reconfigure and adjust its processes. To facilitate the improvement process, the method integrates the synthetic data with advanced analytics and machine learning algorithms as well as and the cause-and-effect relationships. This synergy between human expertise and technological advancements represents a compelling path towards a truly optimized smart factory of the future.
Production planning and control are characterized by unplanned events or so-called turbulences. Turbulences can be external, originating outside the company (e.g., delayed delivery by a supplier), or internal, originating within the company (e.g., failures of production and intralogistics resources). Turbulences can have far reaching consequences for companies and their customers, such as delivery delays due to process delays. For target-optimized handling of turbulences in production, forecasting methods incorporating process data in combination with the use of existing flexibility corridors of flexible production systems offer great potential. Probabilistic, data-driven forecasting methods allow determining the corresponding probabilities of potential turbulences. However, a parallel application of different forecasting methods is required to identify an appropriate one for the specific application. This requires a large database, which often is unavailable and, therefore, must be created first. A simulation-based approach to generate synthetic data is used and validated to create the necessary database of input parameters for the prediction of internal turbulences. To this end, a minimal system for conducting simulation experiments on turbulence scenarios was developed and implemented. A multi-method simulation of the minimal system synthetically generates the required process data, using agent-based modeling for the autonomously controlled system elements and event-based modeling for the stochastic turbulence events. Based on this generated synthetic data and the variation of the input parameters in the forecast, a comparative study of data-driven probabilistic forecasting methods was conducted using a data analytics tool. Forecasting methods of different types (including regression, Bayesian models, nonlinear models, decision trees, ensemble, deep learning) were analyzed in terms of prediction quality, standard deviation, and computation time. This resulted in the identification ofappropriate forecasting methods, and required input parameters for the considered turbulences.
The fifth generation of mobile communication (5G) is a wireless technology developed to provide reliable, fast data transmission for industrial applications, such as autonomous mobile robots and connect cyber-physical systems using Internet of Things (IoT) sensors. In this context, private 5G networks enable the full performance of industrial applications built on dedicated 5G infrastructures. However, emerging wireless communication technologies such as 5G are a complex and challenging topic for training in learning factories, often lacking physical or visual interaction. Therefore, this paper aims to develop a real-time performance monitoring system of private 5G networks and different industrial 5G devices to visualise the performance and impact factors influencing 5G for students and future connectivity experts. Additionally, this paper presents the first long-term measurements of private 5G networks and shows the performance gap between the actual and targeted performance of private 5G networks.
Most Question-answering (QA) systems rely on training data to reach their optimal performance. However, acquiring training data for supervised systems is both time-consuming and resource-intensive. To address this, in this paper, we propose TFCSG, an unsupervised similar question retrieval approach that leverages pre-trained language models and multi-task learning. Firstly, topic keywords in question sentences are extracted sequentially based on a latent topic-filtering algorithm to construct unsupervised training corpus data. Then, the multi-task learning method is used to build the question retrieval model. There are three tasks designed. The first is a short sentence contrastive learning task. The second is the question sentence and its corresponding topic sequence similarity judgment task. The third is using question sentences to generate their corresponding topic sequence task. The three tasks are used to train the language model in parallel. Finally, similar questions are obtained by calculating the cosine similarity between sentence vectors. The comparison experiment on public question datasets that TFCSG outperforms the comparative unsupervised baseline method. And there is no need for manual marking, which greatly saves human resources.
Since its first publication in 2015, the learning factory morphology has been frequently used to design new learning factories and to classify existing ones. The structuring supports the concretization of ideas and promotes exchange between stakeholders.
However, since the implementation of the first learning factories, the learning factory concept has constantly evolved.
Therefore, in the Working Group "Learning Factory Design" of the International Association of Learning Factories, the existing morphology has been revised and extended based on an analysis of the trends observed in the evolution of learning factory concepts. On the one hand, new design elements were complemented to the previous seven design dimensions, and on the other hand, new design dimensions were added. The revised version of the morphology thus provides even more targeted support in the design of new learning factories in the future.
The market for indoor positioning systems for a variety of applications has grown strongly in recent years. A wide range of systems is available, varying considerably in terms of accuracy, price and technology used. The suitability of the systems is highly dependent on the intended application. This paper presents a concept to use a single low-cost PTZ camera in combination with fiducial markers for indoor position and orientation determination. The intended use case is to capture a plant layout consisting of position, orientation and unique identity of individual facilities. Important factors to consider for the selection of a camera have been identified and the transformation of the marker pose in camera coordinates into a selectable plant coordinate system is described. The concept is illustrated by an exemplary practical implementation and its results.
The increase in product variance and shorter product lifecycles result in higher production ramp-up frequencies and promote the usage of mixed-model lines. The ramp-up is considered a critical step in the product life cycle and in the automotive industry phases of the ramp-up are often executed on separated production lines (pilot lines) or factories (pilot plants) to verify processes and to qualify employees without affecting the production of other products in the mixed-model line. The required financial funds for planning and maintaining dedicated pilot lines prevent small and medium-sized enterprises (SMEs) from the application. Hence, SMEs require different tools for piloting and training during the production ramp-up. Learning islands on which employees can be trained through induced and autonomous learning propose a solution. In this work, a concept for the development and application which contains the required organization, activities, and materials is developed through expert interviews. The results of a case study application with a medium-sized automotive manufacturer show that learning islands are a viable tool for employee qualification and process verification during the ramp-up of mixed-model lines.
The presented research is dedicated to estimation of the correlation between the level of renewable energy sources and the costs of congestion management in electric networks in selected European countries. Data of six countries in North-West European area (Italy, Spain, Germany, France, Poland and Austria) were investigated. Factors considered included grid congestion costs including re-dispatching costs as well as countertrading costs, gross electricity generation, installed capacity of electric generating facilities, installed capacity of electric non-dispatchable renewable energy sources and total electricity consumption. Special attention is paid to the share of renewable energy sources. It is found that the grid congestion costs are not clearly affected by penetration of non-dispatchable renewables in all the analysed countries and therefore a clear mathematical correlation cannot no be extrapolated with the available data. The results of this research show in general a loose dependency of the grid congestion costs on the penetration of renewables and a strong dependency on the total electrical consumption of the country.
Measuring cardiorespiratory parameters in sleep, using non-contact sensors and the Ballistocardiography technique has received much attention due to the low-cost, unobtrusive, and non-invasive method. Designing a user-friendly, simple-to-use, and easy-to-deployment preserving less error-prone remains open and challenging due to the complex morphology of the signal. In this work, using four forcesensitive resistor sensors, we conducted a study by designing four distributions of sensors, in order to simplify the complexity of the system by identifying the region of interest for heartbeat and respiration measurement. The sensors are deployed under the mattress and attached to the bed frame without any interference with the subjects. The four distributions are combined in two linear horizontal, one linear vertical, and one square, covering the influencing region in cardiorespiratory activities. We recruited 4 subjects and acquired data in four regular sleeping positions, each for a duration of 80 seconds. The signal processing was performed using discrete wavelet transform bior 3.9 and smooth level of 4 as well as bandpass filtering. The results indicate that we have achieved the mean absolute error of 2.35 and 4.34 for respiration and heartbeat, respectively. The results recommend the efficiency of a triangleshaped structure of three sensors for measuring heartbeat and respiration parameters in all four regular sleeping positions.