Refine
Document Type
- Conference proceeding (220)
- Journal article (90)
- Book chapter (18)
- Doctoral Thesis (6)
- Book (5)
- Patent / Standard / Guidelines (5)
- Anthology (2)
Language
- English (346) (remove)
Is part of the Bibliography
- yes (346)
Institute
- Technik (346) (remove)
Publisher
- IEEE (141)
- Springer (29)
- Elsevier (25)
- VDE Verlag (23)
- Association for Computing Machinery (8)
- MDPI (7)
- SciTePress (6)
- European Association for the Development of Renewable Energy, Environment and Power Quality (5)
- Dnipro University of Technology (4)
- EMW (4)
On-chip metallization, especially in modern integrated BCD technologies, is often subject to high current densities and pronounced temperature cycles due to heat dissipation from power switches like LDMOS transistors. This paper continues the work on a sensor concept where small sense lines are embedded in the metallization layers above the active area of a switching LDMOS transistor. The sensors show a significant resistance change that correlates with the number of power cycles. Furthermore, influences of sense line layer, geometry and the dissipated energy are shown. In this paper, the focus lies on a more detailed analysis of the observed change in sense line resistance.
Lithographical hotspot (LH) detection using deep learning (DL) has received much attention in the recent years. It happens mainly due to the facts the DL approach leads to a better accuracy over the traditional, state-of-the-art programming approaches. The purpose of ths study is to compare existing data augmentation (DA) techniques for the integrated circuit (IC) mask data using DL methods. DA is a method which refers to the process of creating new samples similar to the training set, thereby helping to reduce the gap between classes as well as improving the performance of the DL system. Experimental results suggest that the DA methods increase overall DL models performance for the hotspot detection tasks.
Bionic optimisation is one of the most popular and efficient applications of bionic engineering. As there are many different approaches and terms being used, we try to come up with a structuring of the strategies and compare the efficiency of the different methods. The methods mostly proposed in literature may be classified into evolutionary, particle swarm and artificial neural net optimisation. Some related classes have to be mentioned as the non-sexual fern optimisation and the response surfaces, which are close to the neuron nets. To come up with a measure of the efficiency that allows to take into account some of the published results the technical optimisation problems were derived from the ones given in literature. They deal with elastic studies of frame structures, as the computing time for each individual is very short. General proposals, which approach to use may not be given. It seems to be a good idea to learn about the applicability of the different methods at different problem classes and then do the optimisation according to these experiences. Furthermore in many cases there is some evidence that switching from one method to another improves the performance. Finally the identification of the exact position of the optimum by gradient methods is often more efficient than long random walks around local maxima.
Perforations of the tympanic membrane (TM) can occur as a result of injury or inflammation of the middle ear. These perforations can lead to conductive hearing loss (HL), where in some cases the magnitude of HL exceeds that attributable to the observed TM perforation alone. We aim with this study to better understand the effects of location and size of TM perforations on the sound transmitting properties of the middle ear.
The middle ear transfer function (METF) of six human temporal bones (TB; freshly frozen specimen of body donors) were compared before and after perforation of the TM at different locations (anterior or posterior lower quadrant) and of different sizes (1mm, ¼ of the TM, ½ of the TM, and full ablation). The
METF were correlated with a Finite Element (FE) model of the middle ear, in which similar alterations were simulated.
The measured and simulated FE model METFs exhibited frequency and perforation size dependent amplitude losses at all locations and severities. In direct comparison, posterior TM perforations affected the transmission properties to a larger degree than perforations of the anterior quadrant. This could possibly be caused by an asymmetry of the TM, where the malleus-incus complex rotates and results in larger deflections in the posterior TM half than in the anterior TM half. The FE model of the TM with a sealed cavity suggest that small perforations result in a decrease of TM rigidity and thus to an increase in oscillation amplitude of the TM, mostly above 1 kHz.
The location and size of TM perforations influence the METF in a reproducible way. Correlating our data with the FE model could help to better understand the pathologic mechanisms of middle-ear diseases. If small TM perforations with uncharacteristically significant HL are observed in daily clinical practice, additional middle ear pathologies should be considered. Further investigations on the loss of TM pretension due to perforations may be informative.
Reconstructing 3D face shape from a single 2D photograph as well as from video is an inherently ill-posed problem with many ambiguities. One way to solve some of the ambiguities is using a 3D face model to aid the task. 3D morphable face models (3DMMs) are amongst the state of the art methods for 3D face reconstruction, or so called 3D model fitting. However, current existing methods have severe limitations, and most of them have not been trialled on in-the-wild data. Current analysis-by- synthesis methods form complex non linear optimisation processes, and optimisers often get stuck in local optima. Further, most existing methods are slow, requiring in the order of minutes to process one photograph.
This thesis presents an algorithm to reconstruct 3D face shape from a single image as well as from sets of images or video frames in real-time. We introduce a solution for linear fitting of a PCA shape identity model and expression blendshapes to 2D facial landmarks. To improve the accuracy of the shape, a fast face contour fitting algorithm is introduced. These different components of the algorithm are run in iteration, resulting in a fast, linear shape-to- landmarks fitting algorithm. The algorithm, specifically designed to fit to landmarks obtained from in-the-wild images, by tackling imaging conditions that occur in in-the-wild images like facial expressions and the mismatch of 2D–3D contour correspondences, achieves the shape reconstruction accuracy of much more complex, nonlinear state of the art methods, while being multiple orders of magnitudes faster.
Second, we address the problem of fitting to sets of multiple images of the same person, as well as monocular video sequences. We extend the proposed shape-to-landmarks fitting to multiple frames by using the knowledge that all images are from the same identity. To recover facial texture, the approach uses texture from the original images, instead of employing the often-used PCA albedo model of a 3DMM. We employ an algorithm that merges texture from multiple frames in real-time based on a weighting of each triangle of the reconstructed shape mesh.
Last, we make the proposed real-time 3D morphable face model fitting algorithm available as open-source software. In contrast to ubiquitous available 2D-based face models and code, there is a general lack of software for 3D morphable face model fitting, hindering a widespread adoption. The library thus constitutes a significant contribution to the community.
We present a fully automatic approach to real-time 3D face reconstruction from monocular in-the-wild videos. We use a 3D morphable face model to obtain a semi-dense shape and combine it with a fast median-based super-resolution technique to obtain a high-fidelity textured 3D face model. Our system does not need prior training and is designed to work in uncontrolled scenarios.
We present a fully automatic approach to real-time 3D face reconstruction from monocular in-the-wild videos. With the use of a cascaded-regressor-based face tracking and a 3D morphable face model shape fitting, we obtain a semidense 3D face shape. We further use the texture information from multiple frames to build a holistic 3D face representation from the video footage. Our system is able to capture facial expressions and does not require any person specific training. We demonstrate the robustness of our approach on the challenging 300 Videos in the Wild (300- VW) dataset. Our real-time fitting framework is available as an open-source library at http://4dface.org.
In this work we investigate the behavior of MIS- and Schottky-gate AlGaN/GaN HEMTs under high-power pulsestress. A special setup capable of applying pulses of constant power is used to evaluate the electro-thermal response in different operating points. For both types of devices, the time to failure was found to decrease with increasing drain-source voltage. Overall, the Schottky-gate device displays a higher pulse robustness. The pulse withstand time of the MIS-gate device is limited by the occurrence of a thermal instability at approximately 240°C while the Schottky-gate device displays a rapid increase of the gate leakage current prior to failure. The mechanism responsible for this gate current is further investigated by static and transient temperature measurements and yielded activation energies of 0.6 eV and 0.84 eV.
For optimization of production processes and product quality, often knowledge of the factors influencing the process outcome is compulsory. Thus, process analytical technology (PAT) that allows deeper insight into the process and results in a mathematical description of the process behavior as a simple function based on the most important process factors can help to achieve higher production efficiency and quality. The present study aims at characterizing a well-known industrial process, the transesterification reaction of rapeseed oil with methanol to produce fatty acid methyl esters (FAME) for usage as biodiesel in a continuous micro reactor set-up. To this end, a design of experiment approach is applied, where the effects of two process factors, the molar ratio and the total flow rate of the reactants, are investigated. The optimized process target response is the FAME mass fraction in the purified nonpolar phase of the product as a measure of reaction yield. The quantification is performed using attenuated total reflection infrared spectroscopy in combination with partial least squares regression. The data retrieved during the conduction of the DoE experimental plan were used for statistical analysis. A non-linear model indicating a synergistic interaction between the studied factors describes the reactor behavior with a high coefficient of determination (R²) of 0.9608. Thus, we applied a PAT approach to generate further insight into this established industrial process.
Analog integrated circuit sizing is notoriously difficult to automate due to its complexity and scale; thus, it continues to heavily rely on human expert knowledge. This work presents a machine learning-based design automation methodology comprising pre-defined building blocks such as current mirrors or differential pairs and pre-computed look-up tables for electrical characteristics of primitive devices. Modeling the behavior of primitive devices around the operating point with neural networks combines the speed of equation-based methods with the accuracy of simulation-based approaches and, thereby, brings quality of life improvements for analog circuit designers using the gm/Id method. Extending this procedural automation method for human design experts, we present a fully autonomous sizing approach. Related work shows that the convergence properties of conventional optimization approaches improve significantly when acting in the electrical domain instead of the geometrical domain. We, therefore, formulate the circuit sizing task as a sequential decision-making problem in the alternative electrical design space. Our automation approach is based entirely on reinforcement learning, whereby abstract agents learn efficient design space navigation through interaction and without expert guidance. These agents’ learning behavior and performance are evaluated on circuits of varying complexity and different technologies, showing both the feasibility and portability of the work presented here.
We have seen that bionic optimization can be a powerful tool when applied to problems with non-trivial landscapes of goals and restrictions. This, in turn, led us to a discussion of useful methodologies for applying this optimization to real problems. On the other hand, it must be stated that each optimization is a time consuming process as soon as the problem expands beyond a small number of free parameters related to simple parabolic responses. Bionic optimization is not a quick approach to solving complex questions within short times. In some cases it has the potential to fail entirely, either by sticking to local maxima or by random exploration of the parameter space without finding any promising solutions. The following sections present some remarks on the efficiency and limitations users must be aware of. They aim to increase the knowledge base of using and encountering bionic optimization. But they should not discourage potential users from this promising field of powerful strategies to find good or even the best possible designs.
With the rapid development of globalization, the demand for translation between different languages is also increasing. Although pre-training has achieved excellent results in neural machine translation, the existing neural machine translation has almost no high-quality suitable for specific fields. Alignment information, so this paper proposes a pre-training neural machine translation with alignment information via optimal transport. First, this paper narrows the representation gap between different languages by using OTAP to generate domain-specific data for information alignment, and learns richer semantic information. Secondly, this paper proposes a lightweight model DR-Reformer, which uses Reformer as the backbone network, adds Dropout layers and Reduction layers, reduces model parameters without losing accuracy, and improves computational efficiency. Experiments on the Chinese and English datasets of AI Challenger 2018 and WMT-17 show that the proposed algorithm has better performance than existing algorithms.
Purpose
The purpose of this study is to examine private households’ preferences for service bundles in the German energy market.
Design/methodology/approach
This investigation is based on survey data collected from 3,663 customers of seven mainly municipal energy suppliers in the German energy market. The data set was analyzed via a binary logistic regression model to identify the most prospective customers and their preferences regarding bundles of energy services.
Findings
The results indicate that potential adopters of energy-related service bundles have greater prior knowledge about service bundles; place higher importance on simplified handling, flat rates and long price guarantees; prefer to purchase a service bundle from an energy supplier; live in urban areas and have a gas tariff; are both less likely to have a green electricity tariff and to support the German energy transition; have a greater intention to purchase a smart home product; are less likely to already be prosumers; and prefer customer centers and social media as communication channels with energy providers.
Practical implications
This paper offers several implications for decision-makers in developing marketing strategies for bundled offerings in a highly competitive energy market.
Originality/value
This paper contributes to the sparse research on service bundles in the energy sector, despite the growing interest of energy suppliers and consumers in this topic. It expands the research focusing on the telecommunications sector.
In clothing e-commerce, the challenge of optimally recommending clothing that suits a user’s unique characteristics remains a pressing issue. Many platforms simply recommend best-selling or popular clothing, without taking into account important attributes like user’s face color, pupil color, face shape, age, etc. To solve this problem, this paper proposes a personalized clothing recommendation algorithm that incorporates the established 4-Season Color System and user-specific biological characteristics. Firstly, the attributes and colors of clothing are classified by Fnet network, that can learn disjoint label combinations and mitigate the issue of excessive labels. Secondly, on the basis of the 4-Season Color System, the user’s face color model is trained by combined MobileNetV3_DTL, which ensures the model’s generalization and improves the training speed. Thirdly, user’s face shape and age are divided into different categories by an Inception network. Finally, according to the users’ face color, age, face shape and other information, personalized clothing is recommended in a coarse-to-fine manner. Experiments on five datasets demonstrate that the algorithm proposed in this paper achieves state-of-the-art results.
With the continuous development of economy, consumers pay more attention to the demand for personalization clothing. However, the recommendation quality of the existing clothing recommendation system is not enough to meet the user’s needs. When browsing online clothing, facial expression is the salient information to understand the user’s preference. In this paper, we propose a novel method to automatically personalize clothing recommendation based on user emotional analysis. Firstly, the facial expression is classified by multiclass SVM. Next, the user’s multi-interest value is calculated using expression intensity that is obtained by hybrid RCNN. Finally, the multi-interest value is fused to carry out personalized recommendation. The experimental results show that the proposed method achieves a significant improvement over other algorithms.
We presented our robot framework and our efforts to make face analysis more robust towards self-occlusion caused by head pose. By using a lightweight linear fitting algorithm, we are able to obtain 3D models of human faces in real-time. The combination of adaptive tracking and 3D face modelling for the analysis of human faces is used as a basis for further research on human-machine interaction on our SCITOS robot platform.
The loss contribution of a 2.3kW synchronous GaN-HEMT boost converter for an input voltage of 250V and an output voltage of 500V was analyzed. A simulation model which consists of two parts is introduced. First, a physics-based model is used to determine the switching losses. Then, a system simulation is applied to calculate the losses of the specific elements. This approach allows a fast and accurate system evaluation as required for further system optimization.
In this work, a hard- and a zero-voltage turn-on switching converter are compared. Measurements were performed to verify the simulation model, showing a good agreement. A peak efficiency of 99% was achieved for an output power of 1.4kW. Even with an output power above 400W, it was possible to obtain a system efficiency exceeding 98 %.
We present the results of an extensive characterization of the performance and stability of a third-order continuous-time delta-sigma modulator with active coefficient error compensation. Using our previously published coefficient tuning technique, process variation induced R-C time-constant (TC) errors in the forward signal path can be compensated indirectly using continuously tunable DACs in the feedback path. To validate our technique experimentally with a range of real TC variations, we designed a modulator with discretely configurable integration capacitor arrays in a 0.35-μm CMOS process. We configured the capacitors of the fabricated device for a range of total TC variations from -28.4 % to +19.3 % and measured the signal-to-noise ratio (SNR) as a function of the input amplitude before and after compensating the variations electrically using the feedback DACs. The results show that our tuning technique is capable of restoring the desired nominal modulator performance over the entire parameter variation range, including the system’s nominal maximum stable amplitude (MSA).
While digital IC design is highly automated, analog circuits are still handcrafted in a time-consuming, manual fashion today. This paper introduces a novel Parameterized Circuit Description Scheme (PCDS) for the development of procedural analog schematic generators as parameterized circuits. Circuit designers themselves can use PCDS to create circuit automatisms which capture valuable expert knowledge, offer full topological flexibility, and enhance the re-use of well-established topologies. The generic PCDS concept has been successfully implemented and employed to create parameterized circuits for a broad range of use cases. The achieved results demonstrate the efficiency of our PCDS approach and the potential of parameterized circuits to increase automation in circuit design, also to benefit physical design by promoting the common schematic-driven-layout flow, and to enhance the applicability of circuit synthesis approaches.
Nowadays, the demand for a MEMS development/design kit (MDK) is even more in focus than ever before. In order to achieve a high quality and cost effectiveness in the development process for automotive and consumer applications, an advanced design flow for the MEMS (micro electro mechanical systems) element is urgently required. In this paper, such a development methodology and flow for parasitic extraction of active semiconductor devices is presented. The methodology considers geometrical extraction and links the electrically active pn junctions to SPICE standard library models and subsequently extracts the netlist. An example for a typical pressure sensor is presented and discussed. Finally, the results of the parasitic extraction are compared with fabricated devices in terms of accuracy and capability.
Socially interactive robots with human-like speech synthesis and recognition, coupled with humanoid appearance, are an important subject of robotics and artificial intelligence research. Modern solutions have matured enough to provide simple services to human users. To make the interaction with them as fast and intuitive as possible, researchers strive to create transparent interfaces close to human-human interaction. Because facial expressions play a central role in human-human communication, robot faces were implemented with varying degrees of human-likeness and expressiveness. We propose a way to implement a program that believably animates changing facial expressions and allows to influence them via inter-process communication based on an emotion model. This will can be used to create a screen based virtual face for a robotic system with an inviting appearance to stimulate users to seek interaction with the robot.
This study describes a non-contact measuring and parameter identification procedure designed to evaluate inhomogeneous stiffness and damping characteristics of the annular ligament in the physiological amplitude and frequency range without the application of large static external forces that can cause unnatural displacements of the stapes. To verify the procedure, measurements were first conducted on a steel beam. Then, measurements on an individual human cadaveric temporal bone sample were performed. The estimated results support the inhomogeneous stiffness and damping distribution of the annular ligament and are in a good agreement with the multiphoton microscopy results which show that the posterior-inferior corner of the stapes footplate is the stiffest region of the annular ligament. This method can potentially help to establish a correlation between stiffness and damping characteristics of the annular ligament and inertia properties of the stapes and, thus, help to reduce the number of independent parameters in the model-based hearing diagnosis.
Current clinical practice is often unable to identify the causes of conductive hearing loss in the middle ear with sufficient certainty without exploratory surgery. Besides the large uncertainties due to interindividual variances, only partially understood cause–effect principles are a major reason for the hesitant use of objective methods such as wideband tympanometry in diagnosis, despite their high sensitivity to pathological changes. For a better understanding of objective metrics of the middle ear, this study presents a model that can be used to reproduce characteristic changes in metrics of the middle ear by altering local physical model parameters linked to the anatomical causes of a pathology. A finite-element model is, therefore, fitted with an adaptive parameter identification algorithm to results of a temporal bone study with stepwise and systematically prepared pathologies. The fitted model is able to reproduce well the measured quantities reflectance, impedance, umbo and stapes transfer function for normal ears and ears with otosclerosis, malleus fixation, and disarticulation. In addition to a good representation of the characteristic influences of the pathologies in the measured quantities, a clear assignment of identified model parameters and pathologies consistent with previous studies is achieved. The identification results highlight the importance of the local stiffness and damping values in the middle ear for correct mapping of pathological characteristics and address the challenges of limited measurement data and wide parameter ranges from the literature. The great sensitivity of the model with respect to pathologies indicates a high potential for application in model-based diagnosis.
In this article feedback linearization for control-affine nonlinear systems is extended to systems where linearization is not feasible in the complete state space by combining state feedback linearization and homotopy numerical continuation in subspaces of the phase space where feedback linearization fails. Starting from the conceptual simplicity of feedback linearization, this new method expands the scope of their applicability to irregular systems with poorly expressed relative degree. The method is illustrated on a simple SISO–system and by controlling the speed and the rotor flux linkage in a three phase induction machine.
AI-based prediction and recommender systems are widely used in various industry sectors. However, general acceptance of AI-enabled systems is still widely uninvestigated. Therefore, firstly we conducted a survey with 559 respondents. Findings suggested that AI-enabled systems should be fair, transparent, consider personality traits and perform tasks efficiently. Secondly, we developed a system for the Facial Beauty Prediction (FBP) benchmark that automatically evaluates facial attractiveness. As our previous experiments have proven, these results are usually highly correlated with human ratings. Consequently they also reflect human bias in annotations. An upcoming challenge for scientists is to provide training data and AI algorithms that can withstand distorted information. In this work, we introduce AntiDiscriminationNet (ADN), a superior attractiveness prediction network. We propose a new method to generate an unbiased convolutional neural network (CNN) to improve the fairn ess of machine learning in facial dataset. To train unbiased networks we generate synthetic images and weight training data for anti-discrimination assessments towards different ethnicities. Additionally, we introduce an approach with entropy penalty terms to reduce the bias of our CNN. Our research provides insights in how to train and build fair machine learning models for facial image analysis by minimising implicit biases. Our AntiDiscriminationNet finally outperforms all competitors in the FBP benchmark by achieving a Pearson correlation coefficient of PCC = 0.9601.
The vast majority of state-of-the-art integrated circuits are mixed-signal chips. While the design of the digital parts of the ICs is highly automated, the design of the analog circuitry is largely done manually; it is very time-consuming; and prone to error. Among the reasons generally listed for this is often the attitude of the analog designer. The fact is that many analog designers are convinced that human experience and intuition are needed for good analog design. This is why they distrust the automated synthesis tools. This observation is quite correct, but this is only a symptom of the real problem. This paper shows that this phenomenon is caused by very concrete technical (and thus very rational) issues. These issues lie in the mode of operation of the typical optimization processes employed for the synthesizing tasks. I will show that the dilemma that arises in analog design with these optimizers is the root cause of the low level of automation in analog design. The paper concludes with a review of proposals for automating analog design
In a digitally controlled slope shaping system, reliable detection of both voltage and current slope is required to enable a closed-loop control for various power switches independent of system parameters. In most state-of-the-art works, this is realized by monitoring the absolute voltage and current values. Better accuracy at lower DC power loss is achieved by sensing techniques for a reliable passive detection, which is achieved through avoiding DC paths from the high voltage network into the sensing network. Using a high-speed analog-to-digital converter, the whole waveform of the transient derivative can be stored digitally and prepared for a predictive cycle-by-cycle regulation, without requiring high-precision digital differentiation algorithms. To gain an accurate representation of the voltage and current derivative waveforms, system parasitics are investigated and classified in three sections: (1) component parasitics, which are identified by s-parameter measurements and extraction of equivalent circuit models, (2) PCB design issues related to the sensing circuit, and (3) interconnections between adjacent boards.
The contribution of this paper is an optimized sensing network on the basis of the experimental study supporting fast transition slopes up to 100 V/ns and 1 A/ns and beyond, making the sensing technique attractive for slope shaping of fast switching devices like modern generation IGBTs, CoolMOSTM and SiC mosfets. Measurements of the optimized dv/dt and di/dt setups are demonstrated for a hard switched IGBT power stage.
To prevent high buildings in endangered zones suffering from seismic attack, TMD are applied successfully. In many applications the dampers are placed along the height of the edifice to reduce the damage during the earthquake. The dimensioning of TMD is a multidimensional optimisation problem with many local maxima. To find the absolute best or a very good design, advanced optimisation strategies have to be applied. Bionic optimization proposes different methods to deal with such tasks but requires many repeated studies of the buildings and dampers design. To improve the speed of the analysis, the authors propose a reduced model of the building including the dampers. A series of consecutive generations shows a growing capacity to reduce the impact of an earthquake on the building. The proposals found help to dimension the dampers. A detailed analysis of the building under earthquake loading may yield an efficient design.
This paper presents a control strategy for optimal utilization of photovoltaic (PV) generated power in conjunction with an Energy Storage System (ESS). The ESS is specifically designed to be retrofitted into existing PV systems in an end-user application. It can be attached in parallel to the PV system and connects to existing DC/AC inverters. In particular, the study covers the impact such a modification has on the output power of existing PV panels. A distinct degradation of PV output power was found due to the different power characteristics of PV panel and ESS. To overcome such degradation a novel feedback system is proposed. The feedback system continuously modifies the power characteristic of the ESS to match the PV panel and thus achieves optimal power utilization. Impact on PV and power point tracking performance is analyzed. Simulation of the proposed system is performed in MATLAB/Simulink. The results are found to be satisfactory.
This paper discusses the optimal control problem for increasing the energy efficiency of induction machines in dynamic operation including field weakening regime. In an offline procedure optimal current and flux trajectories are determined such that the copper losses are minimized during transient operations. These trajectories are useful for a subsequent online implementation.
In many automotive applications, repetitive selfheating is the most critical operation condition for LDMOS transistors in smart power ICs. This is attributed to thermomechanical stress in the on-chip metallization, which results from the different thermal expansion coefficients of the metal and the intermetal dielectric. After many cycles, the accumulated strain in the metallization can lead to short circuits, thus limiting the lifetime. Increasing the LDMOS size can help to lower peak temperatures and therefore to reduce the stress. The downside of this is a higher cost. Hence, it has been suggested to use resilient systems that monitor the LDMOS metallization and lower the stress once a certain level of degradation is reached. Then, lifetime requirements can be fulfilled without oversizing LDMOS transistors, even though a certain performance loss has to be accepted. For such systems, suitable sensors for metal degradation are required. This work proposes a floating metal line embedded in the LDMOS metallization. The suitability of this approach has been investigated experimentally by test structures and shown to be a promising candidate. The obtained results will be explained by means of numerical thermo-mechanical simulations.
Wave-like differential equations occur in many engineering applications. Here the engineering setup is embedded into the framework of functional analysis of modern mathematical physics. After an overview, the –Hilbert space approach to free Euler–Bernoulli bending vibrations of a beam in one spatial dimension is investigated. We analyze in detail the corresponding positive, selfadjoint differential operators of 4-th order associated to the boundary conditions in statics. A comparison with free string wave swinging is outlined.
On the influence of ground and substrate on the radiation characteristics of planar spiral antennas
(2022)
The unidirectional radiation of spiral antennas mounted on a substrate requires the presence of a ground plane. In this work, we successively illustrate the impact of dielectric material and ground plane on the key metrics of a planar equiangular spiral antenna (PESA). For this purpose, a PESA mounted on several substrates with different dielectric properties and thicknesses is modeled and simulated. We introduce the tertiary current flowing on spiral arms when backed by a ground plane.
Novel design for a coreless printed circuit board transformer realizing high bandwidth and coupling
(2019)
Rogowski coils offer galvanic isolation and can measure alternating currents with a high bandwidth. Coreless printed circuit board (PCB) transformers have been used as an alternative to limit the additional stray inductance if a Rogowski coil can not be attached to the circuit. A new PCB transformer layout is proposed to reduce cost, decrease additional stray inductance, increase the bandwidth of current measurements and simplify the integration into existing designs.
This paper presents a novel multi-modal CNN architecture that exploits complementary input cues in addition to sole color information. The joint model implements a mid-level fusion that allows the network to exploit cross modal interdependencies already on a medium feature-level. The benefit of the presented architecture is shown for the RGB-D image understanding task. So far, state-of-the-art RGB-D CNNs have used network weights trained on color data. In contrast, a superior initialization scheme is proposed to pre-train the depth branch of the multi-modal CNN independently. In an end-to-end training the network parameters are optimized jointly using the challenging Cityscapes dataset. In thorough experiments, the effectiveness of the proposed model is shown. Both, the RGB GoogLeNet and further RGB-D baselines are outperformed with a significant margin on two different tasks: semantic segmentation and object detection. For the latter, this paper shows how to extract object level groundtruth from the instance level annotations in Cityscapes in order to train a powerful object detector.
Service robots need to be aware of persons in their vicinity in order to interact with them. People tracking enables the robot to perceive persons by fusing the information of several sensors. Most robots rely on laser range scanners and RGB cameras for this task. The thesis focuses on the detection and tracking of heads. This allows the robot to establish eye contact, which makes interactions feel more natural.
Developing a fast and reliable pose invariant head detector is challenging. The head detector that is proposed in this thesis works well on frontal heads, but is not fully pose-invariant. This thesis further explores adaptive tracking to keep track of heads that do not face the robot. Finally, head detector and adaptive tracker are combined within a new people tracking framework and experiments show its effectiveness compared to a state-of the-art system.
Motivation
(2016)
Since human beings started to work consciously with their environment, they have tried to improve the world they were living in. Early use of tools, increasing quality of these tools, use of new materials, fabrication of clay pots, and heat treatment of metals: all these were early steps of optimization. But even on lower levels of life than human beings or human society, we find optimization processes. The organization of a herd of buffalos to face their enemies, the coordinated strategies of these enemies to isolate some of the herd’s members, and the organization of bird swarms on their long flights to their winter quarters: all these social interactions are optimized strategies of long learning processes, most of them the result of a kind of collective intelligence acquired during long selection periods.
One of the challenges in condition monitoring systems is the residual life time prediction. This prediction is done based on statistical methods, based on physical knowledge about the considered process or a combination of these approaches. Physical knowledge of the system is a result of long-term experience of process operators. However, it can be gained as well by analyzing appropriately designed process models. The additional benefit of such models is that particular effects and their impact on the process behavior can be analyzed in detail and without plant operation in a shorter time. The current contribution developed in the framework of the research project Model Based Hierarchic Condition Monitoring presents such models for condition monitoring of roller chains. First, already existing high order dynamic models given by nonlinear differential equations of such chains are extended to incorporate effects that occur due to a deterioration of the chain condition. Then, a simple model is developed and compared to the high order model. Based on the two models the change in the process behavior due to a deterioration of the roller chain condition is analyzed to illustrate that these models can be used in future research in the above mentioned research project to better predict the residual life time of the considered roller chains.
Due to the large interindividual variances and the poor optical accessibility of the ear, the specificity of hearing diagnostics today is severely restricted to a certain clinical picture and quantitative assessment. Often only a yes or no decision is possible, which depends strongly on the subjective assessment of the ENT physician. A novel approach, in which objectively obtainable, non invasive audiometric measurements are evaluated using a numerical middle ear model, makes it possible to make the hidden middle ear properties visible and quantifiable. The central topic of this paper is a novel parameter identification algorithm that combines inverse fuzzy arithmetic with an artificial neural network in order to achieve a coherent diagnostic overall picture in the comparison of model and measurement. Its usage is shown at a pathological pattern called malleus fixation where the upper ligament of the malleus is pathologically stiffened.
Model-based hearing diagnosis based on wideband tympanometry measurements utilizing fuzzy arithmetic
(2019)
Today's audiometric methods for the diagnosis of middle ear disease are often based on a comparison of measurements with standard curves that represent the statistical range of normal hearing responses. Because of large inter-individual variances in the middle ear, especially in wideband tympanometry (WBT), specificity and quantitative evaluation are greatly restricted. A new model-based approach could transform today's predominantly qualitative hearing diganostics into a quantitative and tailored, patient-specific diagnosis, by evaluating WBT measurements with the aid of a middle-ear model. For this particular investigation, a finite element model of a human ear was used. It consisted of an acoustic ear canal and a tympanic cavity model, a middle-ear with detailed nonlinear models of the tympanic membrane and annular ligament, and a simplified inner-ear model. This model has made it possible to identify pathologies from measurements, by analyzing the parameters through senstivity studies and parameter clustering. Uncertainties due to the lack of knowledge, subjectivity in numerical implementation and model simplification are taken into account by the application of fuzzy arithmetic. The most confident parameter set can be determined by applying an inverse fuzzy method on the measurement data. The principle and the benefits of this model-based approach are illustrated by the example of a two-mass oscillator, and also by the simulation of the energy absorbance of an ear with malleus fixation, where the parameter changes that are introduced can be determined quantitatively through the system identification.
Simulation models of the middle ear have rarely been used for diagnostic purposes due to their limited predictive ability with respect to pathologies. One big challenge is the large uncertainty and ambiguity in the choice of material parameters of the model.
Typically, the model parameters are determined by fitting simulation results to validation measurements. In a previous study, it was shown that fitting the model parameters of a finite-element model using the middle-ear transfer function and various other measurable output variables from normal ears alone is not sufficient to obtain a good predictive ability of the model on pathological middle-ear conditions. However, the inclusion of validation measurements on one pathological case resulted in a very good predictive ability also for other pathological cases. Although the found parameter set was plausible in all aspects, it was not yet possible to draw conclusions about the uniqueness and the accuracy or the uncertainty of the parameter set.
To answer these questions, statistical solution approaches are used in this study. Using the Monte Carlo method, a large number of plausible model data sets are generated that correctly represent the normal and pathological middle-ear characteristics in terms of various output variables like e.g., impedance, reflectance, umbo, and stapes transfer function. Subsequent principal component analyses (PCA) allow to draw conclusions about correlations, quantitative limits and statistical density of parameter values.
Furthermore, applying inverse PCA yields numerous plausible parameterizations of the middle-ear model, which can be used for data augmentation and training of a neural network which is capable of distinguishing between a normal middle ear and pathologies like otosclerosis, malleus fixation, and disarticulation based on objectively measured quantities like impedance, reflectance, and umbo velocity.
Deep learning-based fabric defect detection methods have been widely investigated to improve production efficiency and product quality. Although deep learning-based methods have proved to be powerful tools for classification and segmentation, some key issues remain to be addressed when applied to real applications. Firstly, the actual fabric production conditions of factories necessitate higher real-time performance of methods. Moreover, fabric defects as abnormal samples are very rare compared with normal samples, which results in data imbalance. It makes model training based on deep learning challenging. To solve these problems, an extremely efficient convolutional neural network, Mobile-Unet, is proposed to achieve the end-to-end defect segmentation. The median frequency balancing loss function is used to overcome the challenge of sample imbalance. Additionally, Mobile-Unet introduces depth-wise separable convolution, which dramatically reduces the complexity cost and model size of the network. It comprises two parts: encoder and decoder. The MobileNetV2 feature extractor is used as the encoder, and then five deconvolution layers are added as the decoder. Finally, the softmax layer is used to generate the segmentation mask. The performance of the proposed model has been evaluated by public fabric datasets and self-built fabric datasets. In comparison with other methods, the experimental results demonstrate that segmentation accuracy and detection speed in the proposed method achieve state-of-the-art performance.
SLAM systems are mainly applied for robot navigation while research on feasibility for motion planning with SLAM for tasks like bin-picking, is scarce. Accurate 3D reconstruction of objects and environments is important for planning motion and computing optimal gripper pose to grasp objects. In this work, we propose the methods to analyze the accuracy of a 3D environment reconstructed using a LSD-SLAM system with a monocular camera mounted onto the gripper of a collaborative robot. We discuss and propose a solution to the pose space conversion problem. Finally, we present several criteria to analyze the 3D reconstruction accuracy. These could be used as guidelines to improve the accuracy of 3D reconstructions with monocular LSD-SLAM and other SLAM based solutions.
This paper presents a machine learning powered, procedural sizing methodology based on pre-computed look-up tables containing operating point characteristics of primitive devices. Several Neural Networks are trained for 90nm and 45nm technologies, mapping different electrical parameters to the corresponding dimensions of a primitive device. This transforms the geometric sizing problem into the domain of circuit design experts, where the desired electrical characteristics are now inputs to the model. Analog building blocks or entire circuits are expressed as a sequence of model evaluations, capturing the sizing strategy and intention of the designer in a procedure, which is reusable across different technology nodes. The methodology is employed for the sizing of two operational amplifiers, and evaluated for two technology nodes, showing the versatility and efficiency of this approach.
Already more than 75 countries pledged to become climate neutral by 2050 and the share of global emissions falling into an emission pricing scheme has steeply increased over the past two years. Even where there are no direct implications for industry (yet), there is a series of subtle pressure points driving an increasing number of companies across the globe to work towards climate neutrality and pledging ambitious carbon reduction goals.
This article sheds light on what the pressure points are, what the subtle triggers and what the underlying considerations, as well as hoped-for benefits of industrial companies to achieve decarbonisation. The observations and ideas presented in this paper are derived from quantitative and qualitative data. The quantitative data was collected within the framework of Energy Efficiency Index of German Industry (EEI). The qualitative data has been collected from interviews in industrial organisations and media documents as well as from professional practice.
Not only societal, work force, supply chain and investor expectations play a large role, but also many strategic considerations which have the potential to make the business more resilient and profitable. Those companies that do not move towards decarbonisation on the other hand may face a costly late mover disadvantage.
This piece uncovers subtle interconnections helping stakeholders from industry and beyond to grasp opportunities and challenges ahead. Taking account of these calls for rethinking economic viability calculations and investment decision making. Doing so may subsequently lead to on-site carbon reduction measures being prioritised to decarbonise effectively.
Understanding the factors that influence the accuracy of visual SLAM algorithms is very important for the future development of these algorithms. So far very few studies have done this. In this paper, a simulation model is presented and used to investigate the effect of the number of scene points tracked, the effect of the baseline length in triangulation and the influence of image point location uncertainty. It is shown that the latter is very critical, while the other all play important roles. Experiments with a well known semi-dense visual SLAM approach are also presented, when used in a monocular visual odometry mode. The experiments show that not including sensor bias and scale factor uncertainty is very detrimental to the accuracy of the simulation results.
The hotspot detection has received much attention in the recent years due to a substantial mismatch between lithography wavelength and semiconductor technology feature size. This mismatch causes diffraction when transferring the layout from design onto a silicon wafer. As a result, open or short circuits (i.e. lithography hotspots) are more likely to be produced. Additionally, increasing numbers of semiconductors devices on a wafer required more time for the lithography hotspot detection analysis. In this work, we propose a fast and accurate solution based on novel artificial neural network (ANN) architecture for precise lithography hotspot detection using a convolution neural network (CNN) adopting a state of-the-art technique. The experimental results showed that the proposed model gained accuracy improvement over current state-of-theart approaches. The final code has been made publicly available.
Contemporary public enterprises differ from their forebears. Today, they are more similar to private enterprises, receiving far more attention than previously, when privatization processes all over the world were in the spotlight. Furthermore, the broad research stream of entrepreneurship has so far neglected the consideration of public enterprises. To set a future research agenda, the author examines the dispersed literature using an integrative and organizing framework to identify major topics and research findings. This paper reviews articles that investigate the entrepreneurship in contemporary public enterprises. Despite the growing scholarly interest globally, this systematic literature review indicates there is no more than a loose connection between the literature streams of public entrepreneurship and corporate entrepreneurship. Specifically, the review shows that the multidimensional concept of entrepreneurial orientation has thus far been ignored, although autonomy plays a significant role in the literature review, namely in the context of the interference of the public owner. It also reveals other essential research gaps, such as the development of a modern theory of public enterprises. The linked research stream of public-sector corporate entrepreneurship offers a broad area of scholarly research and should encourage further investigation.