Open Access

Multi-Dimensional Connectionist Classification is amethod for weakly supervised training of Deep Neural Networksfor segmentation-free multi-line offline handwriting recognition.MDCC applies Conditional Random Fields as an alignmentfunction for this task. We discuss the structure and patterns ofhandwritten text that can be used for building a CRF. Since CRFsare cyclic graphical models, we have to resort to approximateinference when calculating the alignment of multi-line text duringtraining, here in the form of Loopy Belief Propagation. This workconcludes with experimental results for transcribing small multi-line samples from the IAM Offline Handwriting DB which showthat MDCC is a competitive methodology.

Pascal Laube presents machine learning approaches for three key problems of reverse engineering of defective structured surfaces: parametrization of curves and surfaces, geometric primitive classification and inpainting of high-resolution textures. The proposed methods aim to improve the reconstruction quality while further automating the process. The contributions demonstrate that machine learning can be a viable part of the CAD reverse engineering pipeline.

Erfahrungsbericht zum Freistellungssemester im Wintersemester 2019/2020 an der University of Auckland, Neuseeland

Der Tätigkeits- und Erfahrungsbericht gibt einen Überblick über die wesentlichen Inhalte des im Wintersemester 2019/20 absolvierten Forschungssemesters und beinhaltet überdies eine Beschreibung der für die anwendungsbezogene Lehre und Forschung gewonnenen Erfahrungen.

The following sabbatical semester report documents the activities accomplished, the valuable experiences, and the resulting and potential gains for the University of Applied Sciences.

Many countries offer state credit guarantees to support credit-constrained exporters. The policy instrument is commonly justified by governments as a means to mitigating adverse outcomes of financial market frictions for exporting firms. Accumulated returns to the German state credit guarantee scheme deriving from risk-compensating premia have outweighed accumulated losses over the past 60 years. Why do private financial agents not step in and provide insurance given that the state-run program yields positive returns? We argue that costs of risk diversification, liquidity management, and coordination among creditors limit the ability of private financial agents to offer comparable insurance products. Moreover, we suggest that the government’s greater effectiveness in recovering claims in foreign countries endows the state with a cost advantage in dealing with the risks involved in large export projects. We test these hypotheses using monthly firm-level data combined with official transaction-level data on covered exports of German firms and find suggestive evidence that positive effects on trade are due to mitigated financial constraints: State credit guarantees benefit firms that are dependent on external finance, if the value at risk which they seek to cover is large, and at times when refinancing conditions on the private financial market are tight.

Anmerkung zum EuGH, Urteil vom 17.12.2019 - C-465/19, BeckRS 2019, 32923 – „B & L Elektrogeräte GmbH Beck-online Dokument: LMK 2020, 427615

This paper presents a new likelihood-based partitioning method of the measurement set for the extended object probability hypothesis density (PHD) filter framework. Recent work has mostly relied on heuristic partitioning methods that cluster the measurement data based on a distance measure between the single measurements. This can lead to poor filter performance if the tracked extended objects are closely spaced. The proposed method called Stochastic Partitioning (StP) is based on sampling methods and was inspired by a former work of Granström et. al. In this work, the StP method is applied to a Gaussian inverse Wishart (GIW) PHD filter and compared to a second filter implementation that uses the heuristic Distance Partitioning (DP) method. The performance is evaluated in Monte Carlo simulations in a scenario where two objects approach each other. It is shown that the sampling based StP method leads to an improved filter performance compared to DP.

A physics lab-setup has been developed for engineering students in their first year at university. The so-called LabTeamCoaching helps to improve general lab skills, such as preparing an experiment, writing a documentation, using graphs and drawing conclusions. By using a flipped classroom approach, students get better involved than in our former physics labs when we applied classical methods. This approach will be described and an overview of our 10 years of experience using this method will be given.

Das hier vorgestellte Netzoptimierungstool kann dem Verteilnetzbetreiber bei einem Störfall im Netz in Echtzeit eine Lösung zur Steuerung seiner Betriebsmittel vorschlagen. Dadurch kann das bestehende Netz optimal genutzt werden und ein kostenintensiver Netzausbau im Mittel- und Niederspannungsnetz verringert oder sogar verhindert werden. Als Grundlage für den Netzoptimierer dient ein künstliches neuronales Netz (KNN). Zum Training des KNN wurden Störfälle generiert, die auf reellen Erzeugungs- und Lastprofilen aus dem CoSSMic-Projekt basieren [1]. Für jeden Störfall wurde aus allen möglichen und sinnvollen Netzkonfigurationen eine optimierte Netztopologie anhand von Lastflussberechnungen ermittelt. Durch die Variation der Stufenschalter der Transformatoren und der Stellungen aller installierten Schalter im Netz wurde berechnet, wie der Stromfluss gelenkt werden muss, damit keines der Betriebsmittel die zulässigen Belastungsgrenzen mehr überschreitet. Für ein virtuelles Testnetz konnte mit einem trainierten KNN zu 90 Prozent die optimale Lösung des jeweiligen Störfalls erkannt werden. Durch die Anwendung der N-Best Methode konnte die Vorhersagewahrscheinlichkeit auf annähernd 99 Prozent erhöht werden.

We present an innovative decision support system (DSS) for distribution system operators (DSO) based on an artificial neural network (ANN). A trained ANN has the ability to recognize problem patterns and to propose solutions that can be implemented directly in real time grid management. The principle functionality of this ANN based optimizer has been demonstrated by means of a simple virtual electrical grid. For this grid, the trained ANN predicted the solution minimizing the total line power dissipation in 98 percent of the cases considered. In 99 percent of the cases, a valid solution in compliance with the specified operating conditions was found. First ANN tests on a more realistic grid, calibrated with household load measurements, revealed a prediction rate between 88 and 90 percent depending on the optimization criteria. This approach promises a faster, more cost-efficient and potentially secure method to support distribution system operators in grid management.

This paper summarizes the trends in metallization and interconnection technology in the eyes of the participants of the 8th Metallization and Interconnection Workshop. Participants were asked in a questionnaire to share their view on the future development of metallization technology, the kind of metal used for front side metallization and the future development of interconnection technology. The continuous improvement of the screen-printing technology is reflected in the high expected percentage share decreasing from 88% in three years to still 70% in ten years. The dominating front side metal in the view of the participants will be silver with an expected percentage share of nearly 70% in 2029. Regarding interconnection technologies, the experts of the workshop expect new technologies to gain significant technology shares faster. Whereas in three years soldering on busbars is expected to dominate with a percentage share of 71% it will drop in ten years to 35% in the eyes of the participants. Multiwire and shingling technologies are seen to have the highest potential with expected percentage shares of 33% (multiwire) and 16% (shingling) in ten years.

The investigation of stress requires to distinguish between stress caused by physical activity and stress that is caused by psychosocial factors. The behaviour of the heart in response to stress and physical activity is very similar in case the set of monitored parameters is reduced to one. Currently, the differentiation remains difficult and methods which only use the heart rate are not able to differentiate between stress and physical activity, without using additional sensor data input. The approach focusses on methods which generate signals providing characteristics that are useful for detecting stress, physical activity, no activity and relaxation.

The exposure to the light has a great influence on human beings in their everyday life. Various lighting sources produce light that reaches the human eye and influences a rhythmic release of melatonin hormone, that is a sleep promoting factor. Since the development of new technologies provides more control over illuminance, this work uses an IoT based lighting system to set up dim and bright scenarios. A small study has been performed on the influence of illuminance on sleep latency. The system consists of different light bulbs, sensors and a central bridge which are interconnected like a mesh network. Also, a mobile app has been developed, that allows to adjust the lighting in various rooms. With the help of a ferro-electret sensor, like applied in sleep monitoring systems, a subject’s sleep was monitored. The sensor is placed below the mattress and it collects data, which is stored and processed in a cloud or in other alternative locations. The research was conducted on healthy young subjects after being previously exposed to the preconfigured illuminance for at least three hours before bedtime. The results indicate correlation between sleep onset latency and exposure to different illuminance before bedtime. In a dimmed environment, the subject fell asleep in average 28% faster compared to the brighter environment.

This thesis deals with the object tracking problem of multiple extended objects. For instance, this tracking problem occurs when a car with sensors drives on the road and detects multiple other cars in front of it. When the setup between the senor and the other cars is in a such way that multiple measurements are created by each single car, the cars are called extended objects. This can occur in real world scenarios, mainly with the use of high resolution sensors in near field applications. Such a near field scenario leads a single object to occupy several resolution cells of the sensor so that multiple measurements are generated per scan. The measurements are additionally superimposed by the sensor’s noise. Beside the object generated measurements, there occur false alarms, which are not caused by any object and sometimes in a sensor scan, single objects could be missed so that they not generate any measurements. To handle these scenarios, object tracking filters are needed to process the sensor measurements in order to obtain a stable and accurate estimate of the objects in each sensor scan. In this thesis, the scope is to implement such a tracking filter that handles the extended objects, i.e. the filter estimates their positions and extents. In context of this, the topic of measurement partitioning occurs, which is a pre-processing of the measurement data. With the use of partitioning, the measurements that are likely generated by one object are put into one cluster, also called cell. Then, the obtained cells are processed by the tracking filter for the estimation process. The partitioning of measurement data is a crucial part for the performance of tracking filter because insufficient partitioning leads to bad tracking performance, i.e. inaccurate object estimates. In this thesis, a Gaussian inverse Wishart Probability Hypothesis Density (GIW-PHD) filter was implemented to handle the multiple extended object tracking problem. Within this filter framework, the number of objects are modelled as Random Finite Sets (RFSs) and the objects’ extent as random matrices (RM). The partitioning methods that are used to cluster the measurement data are existing ones as well as a new approach that is based on likelihood sampling methods. The applied classical heuristic methods are Distance Partitioning (DP) and Sub-Partitioning (SP), whereas the proposed likelihood-based approach is called Stochastic Partitioning (StP). The latter was developed in this thesis based on the Stochastic Optimisation approach by Granström et al. An implementation, including the StP method and its integration into the filter framework, is provided within this thesis. The implementations, using the different partitioning methods, were tested on simulated random multi-object scenarios and in a fixed parallel tracking scenario using Monte Carlo methods. Further, a runtime analysis was done to provide an insight into the computational effort using the different partitioning methods. It emphasized, that the StP method outperforms the classical partitioning methods in scenarios, where the objects move spatially close. The filter using StP performs more stable and with more accurate estimates. However, this advantage is associated with a higher computational effort compared to the classical heuristic partitioning methods.

Damit das Arbeitsrecht weniger Arbeit macht: - Rechte und Pflichten von Arbeitnehmer und Arbeitgeber - Rechte des Betriebsrats - Was Sie über Kündigungsschutz, Lohnzahlung im Krankheitsfall und Co. wissen sollten