Refine
Year of publication
- 2020 (110) (remove)
Document Type
- Conference Proceeding (33)
- Article (reviewed) (31)
- Article (unreviewed) (9)
- Bachelor Thesis (8)
- Letter to Editor (6)
- Master's Thesis (6)
- Patent (5)
- Contribution to a Periodical (4)
- Report (3)
- Book (1)
- Part of a Book (1)
- Doctoral Thesis (1)
- Moving Images (1)
- Other (1)
Conference Type
- Konferenzartikel (30)
- Konferenz-Abstract (1)
- Konferenz-Poster (1)
- Sonstiges (1)
Language
- English (74)
- German (35)
- Other language (1)
Keywords
Institute
- Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) (110) (remove)
Open Access
- Closed Access (55)
- Open Access (47)
- Bronze (4)
- Gold (2)
- Closed (1)
The number of use cases for autonomous vehicles is increasing day by day especially in commercial applications. One important application of autonomous vehicles can be found within the parcel delivery section. Here, autonomous cars can massively help to reduce delivery efforts and time by supporting the courier actively. One important component of course is the autonomous vehicle itself. Nevertheless, beside the autonomous vehicle, a flexible and secure communication architecture also is a crucial key component impacting the overall performance of such system since it is required to allow continuous interactions between the vehicle and the other components of the system. The communication system must provide a reliable and secure architecture that is still flexible enough to remain practical and to address several use cases. In this paper, a robust communication architecture for such autonomous fleet-based systems is proposed. The architecture provides a reliable communication between different system entities while keeping those communications secure. The architecture uses different technologies such as Bluetooth Low Energy (BLE), cellular networks and Low Power Wide Area Network (LPWAN) to achieve its goals.
Modeling of Random Variations in a Switched Capacitor Circuit based Physically Unclonable Function
(2020)
The Internet of Things (IoT) is expanding to a wide range of fields such as home automation, agriculture, environmental monitoring, industrial applications, and many more. Securing tens of billions of interconnected devices in the near future will be one of the biggest challenges. IoT devices are often constrained in terms of computational performance, area, and power, which demand lightweight security solutions. In this context, hardware-intrinsic security, particularly physically unclonable functions (PUFs), can provide lightweight identification and authentication for such devices. In this paper, random capacitor variations in a switched capacitor PUF circuit are used as a source of entropy to generate unique security keys. Furthermore, a mathematical model based on the ordinary least square method is developed to describe the relationship between random variations in capacitors and the resulting output voltages. The model is used to filter out systematic variations in circuit components to improve the quality of the extracted secrets.
In dieser Bachelorthesis wurde ein Funktionsmuster eines energieautarken elektronischen Türschildes mit einem 7,8” großen E-Paper-Display und NFC-Konfigurationsschnittstelle entwickelt, auf dem per Smartphone-App und NFC einfach Informationen wie Abwesendheitsnachrichten angezeigt werden können. Hierzu wird ein Kommunikationsprotokoll entwickelt, welches die Kommunikation zwischen App und Türschild spezifiziert, und einen Befehlssatz zur Konfiguration des Türschildes bereitstellt. Das System wird aus amorphen Silizium-Solarzellen versorgt und verfügt über einen LiPo-Akku als Energiespeicher. Durch sorgfältiges Hardware- und Software-seitiges Low-Power-Design beträgt die Leistungsaufnahme im Ruhemodus lediglich 1, 5 μW. Bedingt durch den anwenderfreundlichen, jedoch für Low-Power-Designs ungeeigneten Display-Controller, beträgt der Energieverbrauch während eines Updates 300 mW. Trotzdem zeigt sich, dass das System bei einer Zellfläche von knapp 220 cm2 auch bei schlechter Beleuchtung von 10 lx in dunklen Gängen mehrere Türschild-Updates pro Tag bereitstellen kann.
Bei bimodaler Cochlea-Implantat-/Hörgerät-Versorgung kann es aufgrund seitenverschiedener Signalverarbeitung zu einer zeitlich versetzten Stimulation der beiden Modalitäten kommen. Jüngste Studien haben gezeigt, dass durch zeitlichen Abgleich der Modalitäten die Schalllokalisation bei bimodaler Versorgung verbessert werden kann. Um solch einen Abgleich vornehmen zu können, ist die messtechnische Bestimmung der Durchlaufzeit von Hörgeräten erforderlich. Kommerziell verfügbare Hörgerätemessboxen können diese Werte häufig liefern. Die dazu verwendete Signalverarbeitung wird dabei aber oft nicht vollständig offengelegt. In dieser Arbeit wird ein alternativer und nachvollziehbarer Ansatz zum Design eines simplen Messaufbaus basierend auf einem Arduino DUE Mikrocontroller-Board vorgestellt. Hierzu wurde ein Messtisch im 3D-Druck gefertigt, auf welchem Hörgeräte über einen 2-ccm-Kuppler an ein Messmikrofon angeschlossen werden können. Über einen Latenzvergleich mit dem simultan erfassten Signal eines Referenzmikrofons kann die Durchlaufzeit von Hörgeräten bestimmt werden. Frequenzspezifische Durchlaufzeiten werden mittels einer Kreuzkorrelation zwischen Ziel- und Referenzsignal errechnet. Aufnahme, Ausgabe und Speicherung der Signale erfolgt über einen ATMEL SAM3X8E Mikrocontroller, welcher auf dem Arduino DUE-Board verbaut ist. Über eigens entworfene elektronische Schaltungen werden die Mikrofone und der verwendete Lautsprecher angesteuert. Nach Abschluss einer Messung (Messdauer ca. 5 s) werden die Messdaten seriell an einen PC übertragen, auf dem die Datenauswertung mittels MATLAB erfolgt. Erste Validierungen zeigten eine hohe Stabilität der Messergebnisse mit sehr geringen Standardabweichungen im Bereich weniger Mikrosekunden für Pegel zwischen 50 und 75 dB (A). Der Messaufbau wird in laufenden Studien zur Quantifizierung der Durchlaufzeit von Hörgeräten verwendet.
Many different methods, such as screen printing, gravure, flexography, inkjet etc., have been employed to print electronic devices. Depending on the type and performance of the devices, processing is done at low or high temperature using precursor- or particle-based inks. As a result of the processing details, devices can be fabricated on flexible or non-flexible substrates, depending on their temperature stability. Furthermore, in order to reduce the operating voltage, printed devices rely on high-capacitance electrolytes rather than on dielectrics. The printing resolution and speed are two of the major challenging parameters for printed electronics. High-resolution printing produces small-size printed devices and high-integration densities with minimum materials consumption. However, most printing methods have resolutions between 20 and 50 μm. Printing resolutions close to 1 μm have also been achieved with optimized process conditions and better printing technology.
The final physical dimensions of the devices pose severe limitations on their performance. For example, the channel lengths being of this dimension affect the operating frequency of the thin-film transistors (TFTs), which is inversely proportional to the square of channel length. Consequently, short channels are favorable not only for high-frequency applications but also for high-density integration. The need to reduce this dimension to substantially smaller sizes than those possible with today’s printers can be fulfilled either by developing alternative printing or stamping techniques, or alternative transistor geometries. The development of a polymer pen lithography technique allows scaling up parallel printing of a large number of devices in one step, including the successive printing of different materials. The introduction of an alternative transistor geometry, namely the vertical Field Effect Transistor (vFET), is based on the idea to use the film thickness as the channel length, instead of the lateral dimensions of the printed structure, thus reducing the channel length by orders of magnitude. The improvements in printing technologies and the possibilities offered by nanotechnological approaches can result in unprecedented opportunities for the Internet of Things (IoT) and many other applications. The vision of printing functional materials, and not only colors as in conventional paper printing, is attractive to many researchers and industries because of the added opportunities when using flexible substrates such as polymers and textiles. Additionally, the reduction of costs opens new markets. The range of processing techniques covers laterally-structured and large-area printing technologies, thermal, laser and UV-annealing, as well as bonding techniques, etc. Materials, such as conducting, semiconducting, dielectric and sensing materials, rigid and flexible substrates, protective coating, organic, inorganic and polymeric substances, energy conversion and energy storage materials constitute an enormous challenge in their integration into complex devices.
Background: This paper presents a novel approach for a hand prosthesis consisting of a flexible, anthropomorphic, 3D-printed replacement hand combined with a commercially available motorized orthosis that allows gripping.
Methods: A 3D light scanner was used to produce a personalized replacement hand. The wrist of the replacement hand was printed of rigid material; the rest of the hand was printed of flexible material. A standard arm liner was used to enable the user’s arm stump to be connected to the replacement hand. With computer-aided design, two different concepts were developed for the scanned hand model: In the first concept, the replacement hand was attached to the arm liner with a screw. The second concept involved attaching with a commercially available fastening system; furthermore, a skeleton was designed that was located within the flexible part of the replacement hand.
Results: 3D-multi-material printing of the two different hands was unproblematic and inexpensive. The printed hands had approximately the weight of the real hand. When testing the replacement hands with the orthosis it was possible to prove a convincing everyday functionality. For example, it was possible to grip and lift a 1-L water bottle. In addition, a pen could be held, making writing possible.
Conclusions: This first proof-of-concept study encourages further testing with users.
Um die Akzeptanz und Relevanz von Mailings zu steigern, ist es für Unternehmen wichtig, die Kundeninteressen möglichst gezielt anzusprechen. Bereits jetzt wird die E-Mail-Marketing Lösung Inxmail Professional zusammen mit Recommender Systemen eingesetzt, was das Erstellen und Versenden von E-Mails mit personalisierten Produktempfehlungen ermöglicht. Das notwendige Wissen für den Aufbau und den Betrieb dieser Recommender Systeme liegt dabei jedoch bei (externen) Technologie-Partnern, wodurch höhere Kosten und zusätzliche Abhängigkeiten für Inxmail und seine Kunden entstehen.
Mit dieser Arbeit wurde erforscht, was es für das Unternehmen Inxmail bedeutet, mit Open-Source-Software ein eigenständiges Recommender System aufzubauen und in die E-Mail-Marketing Lösung Inxmail Professional zu integrieren. Hierfür wird Inxmail Professional in einem typischen Kontext mit einem angebundenen Onlineshop System betrachtet. Das Recommender System soll anhand der Daten, die es aus dem Onlineshop System bekommt, individuelle Produktempfehlungen berechnen, um diese automatisiert beim Versand für jeden Empfänger abfragen und in die E-Mail integrieren zu können.
Auf Basis des Machine Learning Servers Harness und der integrierten Universal Recommender Engine wurde ein Recommender System aufgebaut.
Für die Integration des Onlineshop Systems Shopware wurde ein Plugin entwickelt, welches das Recommender System über alle relevanten Benutzeraktionen benachrichtigt. Die Inxmail Professional Integration basiert auf einem Webservice, der implementiert wurde, um die individuellen Produktempfehlungen für einen Kunden/Empfänger beim Recommender System abzufragen und mit den Detailinformationen zu den Produkten aus dem Onlineshop aufzubereiten.
Es konnte gezeigt werden, dass auf der Grundlage von Open-Source-Technologien ein eigenständiges Recommender System aufgebaut, evaluiert und in Inxmail Professional mit angebundenem Onlineshop System integriert werden kann. Der Gesamtprozess für die Evaluation des Systems wurde weitestgehend automatisiert, wodurch viele manuelle und zeitintensive Schritte über Steuerungsskripte abgehandelt werden können. Das System erlaubt die reibungslose Verarbeitung großer Datenmengen (> 19.000.000 Events) auf einem einzelnen Server. Die Vorhersagequalität wurde anhand realer Interaktionsdaten aus öffentlichen E-Commerce-Datensätzen mit Offline-Tests gemessen. Dadurch wurde sichergestellt, dass das System in der Lage ist, bei stetig wachsender Eventmenge zuverlässige aufempfehlungen zu berechnen. Die Tests haben gezeigt, dass bei mehreren Millionen Events von mehreren Tausend Benutzern für mehrere Tausend Produkte, 13 % bis 15 % aller tatsächlichen Käufe in den Top-10 der vom System vorgeschlagenen Empfehlungen enthalten waren.
Damit wurde die Grundlage für den selbstständigen Aufbau und Betrieb eines Recommender Systems gelegt, was eine sinnvolle und kostengünstige Alternative zu externen Systemen sein kann.
Da die hohe Anzahl an Steuergeräten in einem Fahrzeug von den unterschiedlichsten Automobilzulieferern entwickelt und produziert werden, ist es den einzelnen Steuergeräte-Herstellern nicht möglich, diese während des Entwicklungsprozesses in einem realen Fahrzeug zu testen. Restbussimulationen, womit Fahrzeugnetzwerke softwaretechnisch nachgebaut werden, schaffen hierbei Abhilfe.
Für die Entwicklung konkurrenzfähiger, effizienter und wirtschaftlicher Steuergerätesoftware wird die Embedded Software in einzelne Module unterteilt. Dieser modulare Prozess ermöglicht das Implementieren der Embedded Softwaremodule in Steuergeräte unterschiedlicher Fahrzeughersteller, sodass es zu Kosteneinsparungen während der Entwicklung und Wartung kommt. Steuergeräte, welche in unterschiedlichen Fahrzeugen zum Einsatz kommen und eine hohe Anzahl an gleichen Softwaremodulen besitzen, werden in sogenannten „Plattformen“ gehandelt.
Im Rahmen dieser Arbeit wird, analog zu der Plattformsoftware der Steuergeräte eine Plattform-Restbussimulation entworfen. Sie stellt dem Softwareentwickler während des kompletten Entwicklungszyklus eine lauffähige Testumgebung zur Verfügung, welche wichtige Steuergeräte eines Fahrzeugnetzwerks nachbildet. So werden in dieser Arbeit Konzepte erstellt und implementiert, welche eine effiziente und intuitive Benutzung der Plattform-Restbussimulation ermöglichen und alle Plattformkunden mit einer einzigen Simulationsumgebung abdecken. Dies führt zu einer zeitlichen Einsparung bei der Implementierung, Verwaltung und Bedienung.
Annotated training data is essential for supervised learning methods. Human annotation is costly and laborsome especially if a dataset consists of hundreds of thousands of samples and annotators need to be hired. Crowdsourcing emerged as a solution that makes it easier to get access to large amounts of human annotators. Introducing paid external annotators however introduces malevolent annotations, both intentional and unintentional. Both forms of malevolent annotations have negative effects on further usage of the data and can be summarized as spam. This work explores different approaches to post-hoc detection of spamming users and which kinds of spam can be detected by them. A manual annotation checking process resulted in the creation of a small user spam dataset which is used in this thesis. Finally an outlook for future improvements of these approaches will be made.
In Unternehmen entstehen beim Prozess der digitalen Transformation stetig neue Anwendungen und Auswertungen. Es wird viel Zeit und Geld in digitale Lösungen investiert, die das Arbeitsumfeld verbessern. Die Plattformen zur Verwaltung und Verteilung dieser werden dabei jedoch häufig vernachlässigt.
Ziel dieser Arbeit ist es eine solche Plattform in Form eines App-Stores für den Standort zu entwickeln. Dabei liegt der Fokus auf der Verbesserung der App-Nutzung, was durch eine nutzerfreundliche Oberfläche für Endanwender und Entwickler erreicht werden soll. Weiterhin ist eine hohe Wartbarkeit der Plattform notwendig, damit sie auch von einem kleinen Team betrieben werden kann.
Zunächst wurde analysiert, welche Architektur und Technologien für die Umsetzung gut geeignet und welche vom Unternehmen vorgegeben sind. Dabei stellte sich heraus, dass eine Microservice-Architektur am besten geeignet ist. Als Technologie stand lediglich das Framework zur Frontendimplementierung zur Wahl, hier war Angular am besten geeignet.
Durch die Verwendung eines iterativen Prozesses konnten bereits zu Beginn die späteren Nutzer in die Entwicklung eingebunden werden. Dies ermöglichte das Ermitteln aller Anforderungen und Entwerfen einer nutzerfreundlichen Oberfläche. Dieses iterative Vorgehen wurde auch während der Implementierung eingesetzt. Dazu wurden mit Testdaten befüllte Versionen bereitgestellt, damit Nutzer frühzeitig Rückmeldung geben konnten.
Die Erfindung betrifft eine Schaltungsanordnung (10) für ein Kraftfahrzeug, mit einer Hochvolt-Batterie (12) zum Speichern von elektrischer Energie, mit wenigstens einer elektrischen Maschine (14) zum Antreiben des Kraftfahrzeugs, mit einem Stromrichter (16), mittels welchem von der Hochvolt-Batterie (12) bereitstellbare Hochvolt-Gleichspannung in Hochvolt-Wechselspannung zum Betreiben der elektrischen Maschine (14) umwandelbar ist, und mit einem Ladeanschluss (20) zum Bereitstellen von elektrischer Energie zum Laden der Hochvolt-Batterie (12), wobei der Stromrichter (16) als ein Drei-Stufen-Stromrichter ausgebildet ist und wenigstens eine einer Phase (u) der elektrischen Maschine (14) zugeordnete Schaltereinheit (46) aufweist, welche zwei in Reihe geschaltete Schaltergruppen (52, 54) umfasst, die jeweils zwei in Reihe geschaltete IGBTs (T11, T12, T13, T14) aufweisen, wobei zwischen den IGBTs (T11, T12) einer der Schaltergruppen (52, 54) ein Anschluss (64) angeordnet ist, welcher direkt mit einer Leitung (34) des Ladeanschlusses (20) elektrisch verbunden ist.
The status quo of PROFINET, a commonly used industrial Ethernet standard, provides no inherent security in its communication protocols. In this thesis an approach for protecting real-time PROFINET RTC messages against spoofing, tampering and optionally information disclosure is specified and implemented into a real-world prototype setup. Therefor authenticated encryption is used, which relies on symmetric cipher schemes. In addition a procedure to update the used symmetric encryption key in a bumpless manner, e.g. without interrupting the real-time communication, is introduced and realized.
The concept for protecting the PROFINET RTC messages was developed in collaboration with a task group within the security working group of PROFINET International. The author of this thesis has also been part of that task group. This thesis contributes by proofing the practicability of the concept in a real-world prototype setup, which consists of three FPGA-based development boards that communicate with each other to showcase bumpless key updates.
To enable a bumpless key update without disturbing the deterministic real-time traffic by dedicated messages, the key update annunciation and status is embedded into the header. By provisioning two key slots, of which only one is in used, while the other is being prepared, a well-synchronized coordinated switch between the receiver and the sender performs the key update.
The developed prototype setup allows to test the concept and builds the foundation for further research and implementation activities, e.g. the impact of cryptographic operations onto the processing time.
Electrolyte-gated thin-film transistors (EGTs) with indium oxide channel, and expected lifetime of three months, enable low-voltage operation (~1 V) in the field of printed electronics (PEs). The channel width of our printed EGTs is varied between 200 and 1000 μm, whereas a channel length between 10 and 100 μm is used. Due to the lack of uniform performance p-type metal oxide semiconductors, n-type EGTs and passive elements are used to design circuits. For logic gates, transistor-resistor logic has been employed so far, but depletion and enhancement-mode EGTs in a transistor-transistor logic boost the circuit performance in terms of delay and signal swing. In this article, the threshold voltage of the EGT, which determines the operation mode, is tuned through sizing of the EGTs channel geometry. The feasibility of both transistor operation modes is demonstrated for logic gates and ring oscillators. An inverter operating at a supply voltage of 1 V shows a maximum gain of 9.6 and a propagation delay time of 0.7 ms, which represents an improvement of ~ 2x for the gain and oscillation frequency, in comparison with the resistor-transistor logic design. Moreover, the power consumption is reduced by 6x.
Oxide semiconductors have the potential to increase the performance of inkjet printed microelectronic devices such as field-effect transistors (FETs), due to their high electron mobilities. Typical metal oxides are n-type semiconductors, while p-type oxides, although realizable, exhibit lower carriermobilities. Therefore, the circuit design based on oxide semiconductors is mostly in n-type logic only. Here we present an inkjet printed pn-diode based on p- and n-type oxide semiconductors.Copper oxide or nickel oxide is used as p-typesemiconductor whereas n-typesemiconductor is realized with indium oxide. Themeasurements show that the pn-diodes operate in the voltage window typical for printed electronics and the emission coefficient is 1.505 and 2.199 for the copper oxide based and nickel oxidebased pn-diode, respectively.Furthermore, a pn-diode model is developed and integrable into a circuit simulator.
In this work a method for the estimation of current slopes induced by inverters operating interior permanent magnet synchronous machines is presented. After the derivation of the estimation algorithm, the requirements for a suitable sensor setup in terms of accuracy, dynamic and electromagnetic interference are discussed. The boundary conditions for the estimation algorithm are presented with respect to application within high power traction systems. The estimation algorithm is implemented on a field programmable gateway array. This moving least-square algorithm offers the advantage that it is not dependent on vectors and therefore not every measured value has to be stored. The summation of all measured values leads to a significant reduction of the required storage units and thus decreases the hardware requirements. The algorithm is designed to be calculated within the dead time of the inverter. Appropriate countermeasures for disturbances and hardware restrictions are implemented. The results are discussed afterwards.
Generative adversarial networks (GANs) provide state-of-the-art results in image generation. However, despite being so powerful, they still remain very challenging to train. This is in particular caused by their highly non-convex optimization space leading to a number of instabilities. Among them, mode collapse stands out as one of the most daunting ones. This undesirable event occurs when the model can only fit a few modes of the data distribution, while ignoring the majority of them. In this work, we combat mode collapse using second-order gradient information. To do so, we analyse the loss surface through its Hessian eigenvalues, and show that mode collapse is related to the convergence towards sharp minima. In particular, we observe how the eigenvalues of the G are directly correlated with the occurrence of mode collapse. Finally, motivated by these findings, we design a new optimization algorithm called nudged-Adam (NuGAN) that uses spectral information to overcome mode collapse, leading to empirically more stable convergence properties.
Generative adversarial networks are the state of the art approach towards learned synthetic image generation. Although early successes were mostly unsupervised, bit by bit, this trend has been superseded by approaches based on labelled data. These supervised methods allow a much finer-grained control of the output image, offering more flexibility and stability. Nevertheless, the main drawback of such models is the necessity of annotated data. In this work, we introduce an novel framework that benefits from two popular learning techniques, adversarial training and representation learning, and takes a step towards unsupervised conditional GANs. In particular, our approach exploits the structure of a latent space (learned by the representation learning) and employs it to condition the generative model. In this way, we break the traditional dependency between condition and label, substituting the latter by unsupervised features coming from the latent space. Finally, we show that this new technique is able to produce samples on demand keeping the quality of its supervised counterpart.
Generative convolutional deep neural networks, e.g. popular GAN architectures, are relying on convolution based up-sampling methods to produce non-scalar outputs like images or video sequences. In this paper, we show that common up-sampling methods, i.e. known as up-convolution or transposed convolution, are causing the inability of such models to reproduce spectral distributions of natural training data correctly. This effect is independent of the underlying architecture and we show that it can be used to easily detect generated data like deepfakes with up to 100% accuracy on public benchmarks. To overcome this drawback of current generative models, we propose to add a novel spectral regularization term to the training optimization objective. We show that this approach not only allows to train spectral consistent GANs that are avoiding high frequency errors. Also, we show that a correct approximation of the frequency spectrum has positive effects on the training stability and output quality of generative networks.
The term attribute transfer refers to the tasks of altering images in such a way, that the semantic interpretation of a given input image is shifted towards an intended direction, which is quantified by semantic attributes. Prominent example applications are photo realistic changes of facial features and expressions, like changing the hair color, adding a smile, enlarging the nose or altering the entire context of a scene, like transforming a summer landscape into a winter panorama. Recent advances in attribute transfer are mostly based on generative deep neural networks, using various techniques to manipulate images in the latent space of the generator.
In this paper, we present a novel method for the common sub-task of local attribute transfers, where only parts of a face have to be altered in order to achieve semantic changes (e.g. removing a mustache). In contrast to previous methods, where such local changes have been implemented by generating new (global) images, we propose to formulate local attribute transfers as an inpainting problem. Removing and regenerating only parts of images, our Attribute Transfer Inpainting Generative Adversarial Network (ATI-GAN) is able to utilize local context information to focus on the attributes while keeping the background unmodified resulting in visually sound results.
Das Ziel dieser Arbeit ist es eine Reihe an Informationen und Erfahrungen zur Verfügung zu stellen, um es der Hochschule Offenburg zu ermöglichen, den Zumi-Roboter für pädagogische Zwecke, speziell für den neu angebotenen Studiengang „Angewandte Künstliche Intelligenz“, einzusetzen. Sie umfasst die Analyse der verbauten Komponenten, Aufschluss über die Bedienoberflächen, die Handhabung des Roboters und Erfahrungsberichte über das Programmieren mit Zumi. Ebenfalls wurden zwei Vorführprogramme konzipiert, welche an Infotagen zur Promotion der Hochschule eingesetzt werden können. Den größten Teil der Arbeit umschließt eine voll ausgearbeitete Laboraufgabe, welche in kommenden Semestern für den bereits angesprochenen Studiengang zum Einsatz kommen und gegen Ende der Arbeit im Detail erläutert wird.
Es wurden verschiedenste Versuche durchgeführt, um die Komponenten zu analysieren und um deren Genauigkeit, Funktionsweise und Verlässlichkeit bewerten zu können.
Wireless communication technologies play a major role to enable megatrends like Internet of Things (IoT) and Industry 4.0. The Narrowband Wireless WAN (NBWWAN) introduced to meet the long range and low power requirements of spatially distributed wireless communication use cases. These networks introduce additional challenges in testing because the network topology and RF characteristics become particularly complex and thus a multitude of different scenarios must be tested. This paper describes the infrastructure for automated testing of radio communication and for systematic measurements of the network performance of NBWWAN.
One of the main requirements of spatially distributed Internet of Things (IoT) solutions is to have networks with wider coverage to connect many low-power devices. Low-Power Wide-Area Networks (LPWAN) and Cellular IoT(cIOT) networks are promising candidates in this space. LPWAN approaches are based on enhanced physical layer (PHY) implementations to achieve long range such as LoRaWAN, SigFox, MIOTY. Narrowband versions of cellular network offer reduced bandwidth and, simplified node and network management mechanisms, such as Narrow Band IoT (NB-IoT) and Long-Term Evolution for Machines (LTE-M). Since the underlying use cases come with various requirements it is essential to perform a comparative analysis of competing technologies. This article provides systematic performance measurement and comparison of LPWAN and NB-IoT technologies in a unified testbed, also discusses the necessity of future fifth generation (5G) LPWAN solutions.
Diese Arbeit umfasst erste Tests und die Inbetriebnahme eines neuen Prüfplatzes bei der QMK-GmbH. Der Prüfplatz selbst soll in der Lage sein, Leistungsshuntwiderstände kalibrieren zu können. Leistungsshuntwiderstände sind meist eher groß und schwer, damit durch viel Material die Wärmeentwicklung kompensiert werden kann. Zudem sind die Kontaktflächen dementsprechend groß, damit der Übergangswiderstand an den Kontaktflächen möglichst gering ist. Der Widerstandswert selbst ist sehr klein. Standardmäßig liegen Widerstände hier im Bereich von 10 bis 100 Ω. Um so kleine Widerstände möglichst genau messen zu können, muss technisch viel Aufwand betrieben werden. In der Regel wird dies mit einer Vierleitermessung realisiert. Leistungsshuntwiderstände werden aber generell mit einem eher hohen Strom im Bereich von 100 bis 10 000 . Mit dem neuen Prüfplatz soll dies auch umgesetzt werden. Die Widerstände sollen mithilfe von hohen Strömen bis 2 kA kalibriert werden, damit der, für den Prüfling, zutreffende Arbeitsbereich unter Berücksichtigung seiner Eigenerwärmung abgebildet werden kann. Für diese Anwendung wurde ein Prüfplatz entwickelt, der 2 kA zur Verfügung stellen kann und mithilfe eines genauen und kalibrierten Referenzwiderstandes den Widerstand des Kalibriergegenstandes ermitteln kann. Würde man den Aufbau messtechnisch beschreiben, so wird durch eine Konstantstromquelle ein Gleichstrom erzeugt, der beide in Reihe geschalteten Widerstände durchströmt. Damit ist der Strom an beiden Widerständen identisch und kann ermittelt werden. An den Widerständen wird gleichzeitig dessen Spannungsabfall gemessen. Mit dem ermittelten Strom kann anschließend über das Ohmsche Gesetz der „unbekannte“ Widerstandswert des Kalibriergegenstandes ermittelt werden. Dieser wird mit dem Sollwert seines Datenblattes verglichen und in einem Protokoll unter Berücksichtigung der eigenen Messunsicherheit bewertet. Die Messergebnisse werden nach der Messung bzw. Kalibrierung in einem Zertifikat zusammen gefasst, und dem Kunden ausgestellt.
Ziel der Arbeit ist es, eine Kalibriereinrichtung zu entwickeln und zu bewerten, die den Richtlinien und Grundlangen der DAkkS entspricht oder zumindest als Grundlage für eine Akkreditierung bei der DAkkS dient. In erster Linie, soll es mit der Kalibriereinrichtung möglich sein, ISO-Kalibrierungen nach der 9001 Norm durchzuführen und zu bewerten.
Printed electronics (PE) is a fast-growing field with promising applications in wearables, smart sensors, and smart cards, since it provides mechanical flexibility, and low-cost, on-demand, and customizable fabrication. To secure the operation of these applications, true random number generators (TRNGs) are required to generate unpredictable bits for cryptographic functions and padding. However, since the additive fabrication process of the PE circuits results in high intrinsic variations due to the random dispersion of the printed inks on the substrate, constructing a printed TRNG is challenging. In this article, we exploit the additive customizable fabrication feature of inkjet printing to design a TRNG based on electrolyte-gated field-effect transistors (EGFETs). We also propose a printed resistor tuning flow for the TRNG circuit to mitigate the overall process variation of the TRNG so that the generated bits are mostly based on the random noise in the circuit, providing a true random behavior. The simulation results show that the overall process variation of the TRNGs is mitigated by 110 times, and the generated bitstream of the tuned TRNGs passes the National Institute of Standards and Technology - Statistical Test Suite. For the proof of concept, the proposed TRNG circuit was fabricated and tuned. The characterization results of the tuned TRNGs prove that the TRNGs generate random bitstreams at the supply voltage of down to 0.5 V. Hence, the proposed TRNG design is suitable to secure low-power applications in this domain.
Printed electronics (PE) enables disruptive applications in wearables, smart sensors, and healthcare since it provides mechanical flexibility, low cost, and on-demand fabrication. The progress in PE raises trust issues in the supply chain and vulnerability to reverse engineering (RE) attacks. Recently, RE attacks on PE circuits have been successfully performed, pointing out the need for countermeasures against RE, such as camouflaging. In this article, we propose a printed camouflaged logic cell that can be inserted into PE circuits to thwart RE. The proposed cell is based on three components achieved by changing the fabrication process that exploits the additive manufacturing feature of PE. These components are optically look-alike, while their electrical behaviors are different, functioning as a transistor, short, and open. The properties of the proposed cell and standard PE cells are compared in terms of voltage swing, delay, power consumption, and area. Moreover, the proposed camouflaged cell is fabricated and characterized to prove its functionality. Furthermore, numerous camouflaged components are fabricated, and their (in)distinguishability is assessed to validate their optical similarities based on the recent RE attacks on PE. The results show that the proposed cell is a promising candidate to be utilized in camouflaging PE circuits with negligible overhead.
Advances in printed electronics (PE) enables new applications, particularly in ultra-low-cost domains. However, achieving high-throughput printing processes and manufacturing yield is one of the major challenges in the large-scale integration of PE technology. In this article, we present a programmable printed circuit based on an efficient printed lookup table (pLUT) to address these challenges by combining the advantages of the high-throughput advanced printing and maskless point-of-use final configuration printing. We propose a novel pLUT design which is more efficient in PE realization compared to existing LUT designs. The proposed pLUT design is simulated, fabricated, and programmed as different logic functions with inkjet printed conductive ink to prove that it can realize digital circuit functionality with the use of programmability features. The measurements show that the fabricated LUT design is operable at 1 V.
Rectifiersare vital electronic circuits for signal and power conversion in various smart sensor applications. The ability to process low input voltage levels, for example, from vibrational energy harvesters is a major challenge with existing passive rectifiers in printed electronics, stemming mainly from the built-in potential of the diode's p-njunction. To address this problem, in this work, we design, fabricate, and characterize an inkjet-printed full-wave rectifier using diode-connected electrolyte-gated thin-film transistors (EGTs). Using both experimental and simulation approaches, we investigate how the rectifier can benefit from the near-zero threshold voltage of transistors, which can be enabled by proper channel geometry setting in EGT technology. The presented circuit can be operated at 1-V input voltage, featuring a remarkably small voltage loss of 140 mV and a cutoff frequency of ~300 Hz. Below the cutoff frequency, more than 2.6-μW dc power is obtained over the load resistances ranging from 5 to 20 kQ. Furthermore, experiments show that the circuit can work with an input amplitude down to 500 mV. This feature makes the presented design highly suitable for a variety of energy-harvesting applications.
Diese Arbeit beschäftigt sich mit der Biomechanik der Halswirbelsäule (HWS) beim Umgang mit dem Smartphone. Die Kräfte, die auf Wirbelkörper, Wirbelgelenke, Bandscheiben, Muskeln und Bänder wirken, werden mit steigendem Flexionswinkel der HWS größer. Die Beschwerden hingegen, welche der Smartphone-Nacken hervorruft, sind meist akut und mit regelmäßiger Bewegung und der Stärkung der Nackenmuskulatur gut zu behandeln. Eine Therapie ist somit auch zur Vorbeugung geeignet. Doch die Langzeitauswirkungen sind nicht außer Acht zu lassen, denn durch die steigenden Nutzungsmöglichkeiten der Smartphones steigt auch der durchschnittliche tägliche Gebrauch stärker an. So wird vor allem die tägliche Bildschirmzeit bei Jugendlichen immer länger. Das aktuell noch akute Krankheitsbild des Smartphone-Nackens, das nur selten einen chronischen Verlauf nimmt und Langzeitschäden verursacht, könnte sich durch fehlende oder zu späte Maßnahmen zu einem größeren chronischen Krankheitsbild entwickeln.
We introduce an open source python framework named PHS-Parallel Hyperparameter Search to enable hyperparameter optimization on numerous compute instances of any arbitrary python function. This is achieved with minimal modifications inside the target function. Possible applications appear in expensive to evaluate numerical computations which strongly depend on hyperparameters such as machine learning. Bayesian optimization is chosen as a sample efficient method to propose the next query set of parameters.
Machine learning (ML) has become highly relevant in applications across all industries, and specialists in the field are sought urgently. As it is a highly interdisciplinary field, requiring knowledge in computer science, statistics and the relevant application domain, experts are hard to find. Large corporations can sweep the job market by offering high salaries, which makes the situation for small and medium enterprises (SME) even worse, as they usually lack the capacities both for attracting specialists and for qualifying their own personnel. In order to meet the enormous demand in ML specialists, universities now teach ML in specifically designed degree programs as well as within established programs in science and engineering. While the teaching almost always uses practical examples, these are somewhat artificial or outdated, as real data from real companies is usually not available. The approach reported in this contribution aims to tackle the above challenges in an integrated course, combining three independent aspects: first, teaching key ML concepts to graduate students from a variety of existing degree programs; second, qualifying working professionals from SME for ML; and third, applying ML to real-world problems faced by those SME. The course was carried out in two trial periods within a government-funded project at a university of applied sciences in south-west Germany. The region is dominated by SME many of which are world leaders in their industries. Participants were students from different graduate programs as well as working professionals from several SME based in the region. The first phase of the course (one semester) consists of the fundamental concepts of ML, such as exploratory data analysis, regression, classification, clustering, and deep learning. In this phase, student participants and working professionals were taught in separate tracks. Students attended regular classes and lab sessions (but were also given access to e-learning materials), whereas the professionals learned exclusively in a flipped classroom scenario: they were given access to e-learning units (video lectures and accompanying quizzes) for preparation, while face-to-face sessions were dominated by lab experiments applying the concepts. Prior to the start of the second phase, participating companies were invited to submit real-world problems that they wanted to solve with the help of ML. The second phase consisted of practical ML projects, each tackling one of the problems and worked on by a mixed team of both students and professionals for the period of one semester. The teams were self-organized in the ways they preferred to work (e.g. remote vs. face-to-face collaboration), but also coached by one of the teaching staff. In several plenary meetings, the teams reported on their status as well as challenges and solutions. In both periods, the course was monitored and extensive surveys were carried out. We report on the findings as well as the lessons learned. For instance, while the program was very well-received, professional participants wished for more detailed coverage of theoretical concepts. A challenge faced by several teams during the second phase was a dropout of student members due to upcoming exams in other subjects.
In the field of neuroprosthetics, the current state-of-the-art method involves controlling the prosthesis with electromyography (EMG) or electrooculography/electroencephalography (EOG/EEG). However, these systems are both expensive and time consuming to calibrate, susceptible to interference, and require a lengthy learning phase by the patient. Therefore, it is an open challenge to design more robust systems that are suitable for everyday use and meet the needs of patients. In this paper, we present a new concept of complete visual control for a prosthesis, an exoskeleton or another end effector using augmented reality (AR) glasses presented for the first time in a proof-of-concept study. By using AR glasses equipped with a monocular camera, a marker attached to the prosthesis is tracked. Minimal relative movements of the head with respect to the prosthesis are registered by tracking and used for control. Two possible control mechanisms including visual feedback are presented and implemented for both a motorized hand orthosis and a motorized hand prosthesis. Since the grasping process is mainly controlled by vision, the proposed approach appears to be natural and intuitive.
Die Erfindung betrifft ein Verfahren zur Steuerung eines Geräts, insbesondere einer Handprothese oder eines Roboterarms, wobei wenigstens ein an oder im Bezug zu dem Gerät positionierter Marker von einer an einer Bedienperson angeordneten Kamera erkannt wird, wobei ab dem Erkennen des wenigstens einen Markers eine vordefinierte Bewegung der Bedienperson zusammen mit der Kamera erkannt wird und zum Auslösen einer entsprechenden Aktion des Geräts verwendet wird, wobei die vordefinierte Bewegung einer Bedienperson in Form eines Sehstrahls mittels Kamera-Tracking erkannt wird. Weiterhin betrifft die Erfindung eine Anordnung aus einem Gerät, insbesondere einer Handprothese oder eines Roboterarms, und einer AR-Brille zur Durchführung eines derartigen Verfahrens.
Restoring hand motion to people experiencing amputation, paralysis, and stroke is a critical area of research and development. While electrode-based systems that use input from the brain or muscle have proven successful, these systems tend to be expensive and di¨cult to learn. One group of researchers is exploring the use of augmented reality (AR) as a new way of controlling hand prostheses. A camera mounted on eyeglasses tracks LEDs on a prosthetic to execute opening and closing commands using one of two different AR systems. One system uses a rectangular command window to control motion: crossing horizontally signals “open” along one direction and “close” in the opposite direction. The second system uses a circular command window: once control is enabled, gripping strength can be controlled by the direction of head motion. While the visual system remains to be tested with patients, its low cost, ease of use, and lack of electrodes make the device a promising solution for restoring hand motion.
A new concept for robust non-invasive optical activation of motorized hand prostheses by simple and non-contactcommands is presented. In addition, a novel approach for aiding hand amputees is shown, outlining significantprogress in thinking worth testing. In this, personalized 3D-printed artificial flexible hands are combined withcommercially available motorized exoskeletons, as they are used e.g. in tetraplegics.
Oesophageal Electrode Probe and Device for Cardiological Treatment and/or Diagnosis (US20200261024)
(2020)
An oesophageal electrode probe for bioimpedance measurement and/or for neurostimulation is provided; a device for transoesophageal cardiological treatment and/or cardiological diagnosis is also provided; a method for the open-loop or closed-loop control of a cardiological catheter ablation device and/or a cardiological, circulatory and/or respiratory support device is also provided. The oesophageal electrode probe comprises a bioimpedance measuring device for measuring the bioimpedance of at least one part of tissue surrounding the oesophageal electrode probe. The bioimpedance device comprises at least one first and one second electrode. The at least one first electrode is arranged on a side of the oesophageal electrode probe facing towards the heart. The at least one second electrode is arranged on a side of the oesophageal electrode probe facing away from the heart. The device comprises the oesophageal electrode probe and a control and/or evaluation device.
本发明涉及一种用于生物阻抗测量和/或用于神经刺激的食道电极探针(10);用于经食道心脏病治疗和/或心脏病诊断的设备(100);以及一种用于控制或调节用于心脏导管消融装置和/或心脏、循环和/或肺支持装置的方法。食道电极探针包括生物阻抗测量装置,用于测量围绕食道电极探针的组织中的至少一部分组织的生物阻抗。生物阻抗装置包括至少一个第一电极和至少一个第二电极,其中至少一个第一电极(12A)布置在食道电极探针的面向心脏的一侧(14)上,并且至少一个第二电极(12B)布置在食道电极探针背离心脏的一侧(16)上。装置(100)包括食道电极探针(10)和控制和/或评估装置(30),其被配置用于从至少一个第一电极(12A)接收第一生物阻抗测量信号并从至少一个第二电极(12B)接收第二生物阻抗测量信号,并对这些信号进行比较,并且在比较的基础上产生控制信号。该控制信号可以是用于控制或调节心脏导管消融装置和/或心脏、循环和/或肺支持装置的信号。
The precise positioning of mobile systems is a prerequisite for any autonomous behavior, in an industrial environment as well as for field robotics. The paper describes the set up for an experimental platform and its use for the evaluation of simultaneous localization and mapping (SLAM) algorithms. Two approaches are compared. First, a local method based on point cloud matching and integration of inertial measurement units is evaluated. Subsequent matching makes it possible to create a three-dimensional point cloud that can be used as a map in subsequent runs. The second approach is a full SLAM algorithm, based on graph relaxation models, incorporating the full sensor suite of odometry, inertial sensors, and 3D laser scan data.
Mit der Implementierung sowie einer anschließenden aussagekräftigen Evaluierung, soll das, visuelle-inertiale Kartierungs- und Lokalisierungssystem maplab analysiert werden. Hierbei basiert die Kartierung bzw. Lokalisierung auf der Detektion von Umgebungsmerkmalen. Neben der Möglichkeit der Kartenerstellung besteht ferner die Option, mehrere Karten zu fusionieren und somit weitreichende Gebiete zu kartieren sowie für weitere Datenauswertungen zu nutzen. Aufgrund der Durchführung und Bewertung der Ergebnisse in unterschiedlichen Anwendungsszenarien zeigt sich, dass maplab besonders zur Kartierung von Räumen bzw. kleinen Gebäudekomplexen geeignet ist. Die Möglichkeit der Kartenfusionierung bietet weiterhin die Option, den Informationsgehalt von Karten zu erhöhen, welches die Effektivität für eine anschließende Lokalisierung steigert. Bei wachsender Kartierungsgröße hingegen zeigt sich jedoch eine Vergrößerung geometrischer Inkonsistenzen.
A novel approach for synchronization and calibration of a camera and an inertial measurement unit (IMU) in the research-oriented visual-inertial mapping-and localization-framework maplab is presented. Mapping and localization are based on detecting different features in the environment. In addition to the possibility of creating single-case maps, the included algorithms allow merging maps to increase mapping accuracy and obtain large-scale maps. Furthermore, the algorithms can be used to optimize the collected data. The preliminary results show that after appropriate calibration and synchronization maplab can be used efficiently for mapping, especially in rooms and small building environments.
Bei dem vorgestellten Ansatz soll der Auftreffpunkt des Pfeils durch die Kreuzkorrelation von Audio-Signalen bestimmt werden. Das Auftreffen des Pfeils erzeugt ein charakteristisches Geräusch, welches von mehreren Mikrofonen in bestimmter Anordnung um die Dartscheibe herum in elektrische Signale umgewandelt wird. Mithilfe der Schallgeschwindigkeit und den Zeitdifferenzen, welche die Schallwelle zu den einzelnen Mikrofonen benötigt soll dann der Auftreffpunkt berechnet werden.
In this work, we evaluate two different image clustering objectives, k-means clustering and correlation clustering, in the context of Triplet Loss induced feature space embeddings. Specifically, we train a convolutional neural network to learn discriminative features by optimizing two popular versions of the Triplet Loss in order to study their clustering properties under the assumption of noisy labels. Additionally, we propose a new, simple Triplet Loss formulation, which shows desirable properties with respect to formal clustering objectives and outperforms the existing methods. We evaluate all three Triplet loss formulations for K-means and correlation clustering on the CIFAR-10 image classification dataset.
Hohe Kosten bei der Annotation von Daten führen dazu, dass datensparsamere Wege zum Erstellen von Modellen gesucht werden. In dieser Arbeit wird ein Lösungsansatz untersucht, der ausgehend von fokussierten Repräsentationen, datensparsame Lösungen für verschiedene Aufgaben finden soll. Durch einen Multi-Task-Learning-Ansatz trägt das Finden einer Repräsentation gleichzeitig zum Lösen einer Aufgabe bei. Durch Ersetzung einer der Aufgaben werden Wissentransfers datensparsam auf die neue Aufgabe durchgeführt. In der erarbeiteten und evaluierten Lösung können Hyperparameter automatisch gefunden werden. Bei Vergleichen von verschiedenen Ansätzen und Datenmengen ist über die Leistung der Netzwerke zu erkennen, dass der Ansatz insbesondere mit weniger Daten bessere Ergebnisse erzielt. Die Ergebnisse dieser Arbeit lassen eine Bereitstellung als Module zu. Die Module werden im Rahmen dieser Arbeit beschrieben. Abgeschlossen wird die Arbeit mit einem Ausblick auf Verbesserungen und Potenziale der Ansätze.
Method for controlling a device, in particular, a prosthetic hand or a robotic arm (US20200327705A1)
(2020)
A method for controlling a device, in particular a prosthetic hand or a robotic arm, includes using an operator-mounted camera to detect at least one marker positioned on or in relation to the device. Starting from the detection of the at least one marker, a predefined movement of the operator together with the camera is detected and is used to trigger a corresponding action of the device. The predefined movement of the operator is detected in the form of a line of sight by means of camera tracking. A system for controlling a device, in particular a prosthetic hand or a robotic arm, includes a pair of AR glasses adapted to detect the at least one marker and to detect the predefined movement of the operator.
In this study, a facile method to fabricate a cohesive ion‐gel based gate insulator for electrolyte‐gated transistors is introduced. The adhesive and flexible ion‐gel can be laminated easily on the semiconducting channel and electrode manually by hand. The ion‐gel is synthesized by a straightforward technique without complex procedures and shows a remarkable ionic conductivity of 4.8 mS cm−1 at room temperature. When used as a gate insulator in electrolyte‐gated transistors (EGTs), an on/off current ratio of 2.24×104 and a subthreshold swing of 117 mV dec−1 can be achieved. This performance is roughly equivalent to that of ink drop‐casted ion‐gels in electrolyte‐gated transistors, indicating that the film‐attachment method might represent a valuable alternative to ink drop‐casting for the fabrication of gate insulators.
Neurostimulation durch Musik
(2020)
Was ist die Musik und wie wirkt sie sich auf den menschlichen Körper aus? Historisch betrachtet wird die Musik als etwas Göttliches aufgefasst, da sie eine äußerst große Wirkung auf die Emotionen des Menschen besitzt. Dieser Effekt wirkt sich auch psychosomatisch aus und kann das Denken und Handeln des Zuhörers beeinflussen und steuern. So lässt sich beispielsweise das Kaufverhalten allein durch die musikalische Begleitung deutlich manipulieren. Selbst die Motivation lässt sich mit passender Vertonung entweder steigern oder reduzieren. In der heutigen Zivilisation begleitet die Musik den Menschen alltäglich und wird zu vielen verschiedenen Zwecken verwendet. Somit ist die musikalische Stimulation als eine Art Psychotherapie zu werten, die häufig gezielt angewendet wird, aber im Beeinflussten unterbewusst stattfindet. Da natürlich immer noch offene Fragen bezüglich der genauen Wirkung von Musik auf das Gehirn bestehen, werden derzeit im Bereich der Neurowissenschaften viele Studien durchgeführt, um dieses Phänomen nachvollziehen zu können.
Multiple Object Tracking (MOT) is a long-standing task in computer vision. Current approaches based on the tracking by detection paradigm either require some sort of domain knowledge or supervision to associate data correctly into tracks. In this work, we present an unsupervised multiple object tracking approach based on visual features and minimum cost lifted multicuts. Our method is based on straight-forward spatio-temporal cues that can be extracted from neighboring frames in an image sequences without superivison. Clustering based on these cues enables us to learn the required appearance invariances for the tracking task at hand and train an autoencoder to generate suitable latent representation. Thus, the resulting latent representations can serve as robust appearance cues for tracking even over large temporal distances where no reliable spatio-temporal features could be extracted. We show that, despite being trained without using the provided annotations, our model provides competitive results on the challenging MOT Benchmark for pedestrian tracking.
Due to the rapidly increasing storage consumption worldwide, as well as the expectation of continuous availability of information, the complexity of administration in today’s data centers is growing permanently. Integrated techniques for monitoring hard disks can increase the reliability of storage systems. However, these techniques often lack intelligent data analysis to perform predictive maintenance. To solve this problem, machine learning algorithms can be used to detect potential failures in advance and prevent them. In this paper, an unsupervised model for predicting hard disk failures based on Isolation Forest is proposed. Consequently, a method is presented that can deal with the highly imbalanced datasets, as the experiment on the Backblaze benchmark dataset demonstrates.
The recent successes and wide spread application of compute intensive machine learning and data analytics methods have been boosting the usage of the Python programming language on HPC systems. While Python provides many advantages for the users, it has not been designed with a focus on multiuser environments or parallel programming - making it quite challenging to maintain stable and secure Python workflows on a HPC system. In this paper, we analyze the key problems induced by the usage of Python on HPC clusters and sketch appropriate workarounds for efficiently maintaining multi-user Python software environments, securing and restricting resources of Python jobs and containing Python processes, while focusing on Deep Learning applications running on GPU clusters.