Multi-view classification with convolutional neural networks. - In: PLOS ONE, ISSN 1932-6203, Bd. 16 (2021), 1, e0245230, insges. 17 S.
https://doi.org/10.1371/journal.pone.0245230
On the design of stable periodic orbits of a triple pendulum on a cart with experimental validation. - In: Automatica, ISSN 0005-1098, Bd. 125 (2021), 109403
In this paper, we enhance an inversion-based control approach towards the stabilization of a periodic orbit of a multi-link triple pendulum on a cart. To this end, a nominal trajectory is obtained by formulating the considered transition problem as a two-point boundary value problem (BVP) in input-output representation. For solvability of the resulting BVP, a setup function is introduced such that additional parameters are provided in the differential equation of the internal dynamics. Based on the linearized dynamics about the nominal trajectory, a linear-quadratic-Gaussian (LQG) controller is implemented to compensate for measurement noise, model uncertainties, and external disturbances. This way we achieve to force a triple pendulum to move along a non-trivial periodic orbit and render it attractive. The high performance and accuracy of our approach is illustrated on an experimental setup.
https://doi.org/10.1016/j.automatica.2020.109403
Pollen analysis using multispectral imaging flow cytometry and deep learning. - In: The new phytologist, ISSN 1469-8137, Bd. 229 (2021), 1, S. 593-606
Pollen identification and quantification are crucial but challenging tasks in addressing a variety of evolutionary and ecological questions (pollination, paleobotany), but also for other fields of research (e.g. allergology, honey analysis or forensics). Researchers are exploring alternative methods to automate these tasks but, for several reasons, manual microscopy is still the gold standard. In this study, we present a new method for pollen analysis using multispectral imaging flow cytometry in combination with deep learning. We demonstrate that our method allows fast measurement while delivering high accuracy pollen identification. A dataset of 426 876 images depicting pollen from 35 plant species was used to train a convolutional neural network classifier. We found the best-performing classifier to yield a species-averaged accuracy of 96%. Even species that are difficult to differentiate using microscopy could be clearly separated. Our approach also allows a detailed determination of morphological pollen traits, such as size, symmetry or structure. Our phylogenetic analyses suggest phylogenetic conservatism in some of these traits. Given a comprehensive pollen reference database, we provide a powerful tool to be used in any pollen study with a need for rapid and accurate species identification, pollen grain quantification and trait extraction of recent pollen.
https://doi.org/10.1111/nph.16882
Advances in engineering research and application : proceedings of the International Conference on Engineering Research and Applications, ICERA 2020. - Cham : Springer International Publishing, 2021. - 1 Online-Ressource (xiv, 886 p.). - (Lecture notes in networks and systems ; Volume 178) ISBN 978-3-030-64719-3
This proceedings book features volumes gathered selected contributions from the International Conference on Engineering Research and Applications (ICERA 2020) organized at Thai Nguyen University of Technology on December 1-2, 2020. The conference focused on the original researches in a broad range of areas, such as Mechanical Engineering, Materials and Mechanics of Materials, Mechatronics and Micromechatronics, Automotive Engineering, Electrical and Electronics Engineering, and Information and Communication Technology. Therefore, the book provides the research community with authoritative reports on developments in the most exciting areas in these fields.
https://doi.org/10.1007/978-3-030-64719-3
Real-time person orientation estimation and tracking using colored point clouds. - In: Robotics and autonomous systems, ISSN 1872-793X, Bd. 135 (2021), 103665, insges. 13 S.
Robustly estimating the orientations of people is a crucial precondition for a wide range of applications. Especially for autonomous systems operating in populated environments, the orientation of a person can give valuable information to increase their acceptance. Given peoples orientations, mobile systems can apply navigation strategies which take peoples proxemics into account or approach them in a human like manner to perform human robot interaction (HRI) tasks. In this paper, we present an approach for person orientation estimation based on computationally efficient features extracted from colored point clouds, formerly used for a two-class person attribute classification. The classification approach has been extended to the continuous domain while treating the problem of orientation estimation in real time. Furthermore, we present an approach for tracking estimated orientations over time using a Bayesian filter. We will show that tracking can increase the accuracy of orientations by up to 3.69˚ on a dataset recorded with a mobile robot. Best results on this highly challenging dataset are achieved with a regression approach for orientation estimation in combination with tracking. The mean angular error of just 16.49˚ proofs the applicability in real-world scenarios.
https://doi.org/10.1016/j.robot.2020.103665
Algorithmen und Datenstrukturen : eine Einführung mit Java
6., überarbeitete und erweiterte Auflage. - Heidelberg : dpunkt.verlag, 2021. - 1 Online-Ressource (xix, 588 Seiten) ISBN 3-96910-066-6
Intro -- I Grundlegende Konzepte -- Vorbemerkungen und Überblick -- Informatik, Algorithmen und Datenstrukturen -- Historischer Überblick: Algorithmen -- Historie von Programmiersprachen und Java -- Grundkonzepte der Programmierung in Java -- Algorithmische Grundkonzepte -- Intuitiver Algorithmusbegriff -- Beispiele für Algorithmen -- Bausteine für Algorithmen -- Pseudocode-Notation für Algorithmen -- Struktogramme -- Rekursion -- Sprachen und Grammatiken -- Begriffsbildung -- Reguläre Ausdrücke -- Backus-Naur-Form (BNF) -- Elementare Datentypen -- Datentypen als Algebren -- Signaturen von Datentypen -- Der Datentyp bool -- Der Datentyp integer -- Felder und Zeichenketten -- Terme -- Bildung von Termen -- Algorithmus zur Termauswertung -- Datentypen in Java -- Primitive Datentypen -- Referenzdatentypen -- Operatoren -- Algorithmenparadigmen -- Überblick über Algorithmenparadigmen -- Applikative Algorithmen -- Terme mit Unbestimmten -- Funktionsdefinitionen -- Auswertung von Funktionen -- Erweiterung der Funktionsdefinition -- Applikative Algorithmen -- Beispiele für applikative Algorithmen -- Imperative Algorithmen -- Grundlagen imperativer Algorithmen -- Komplexe Anweisungen -- Beispiele für imperative Algorithmen -- Das logische Paradigma -- Logik der Fakten und Regeln -- Deduktive Algorithmen -- Weitere Paradigmen -- Genetische Algorithmen -- Neuronale Netze -- Umsetzung in Java -- Ausdrücke und Anweisungen -- Methoden -- Applikative Algorithmen und Rekursion -- Literaturhinweise zum Teil I -- II Algorithmen -- Ausgewählte Algorithmen -- Suchen in sortierten Folgen -- Sequenzielle Suche -- Binäre Suche -- Sortieren -- Sortieren: Grundbegriffe -- Sortieren durch Einfügen -- Sortieren durch Selektion -- Sortieren durch Vertauschen: BubbleSort -- Sortieren durch Mischen: MergeSort -- QuickSort -- Sortieren durch Verteilen: RadixSort.
https://ebookcentral.proquest.com/lib/kxp/detail.action?docID=6379309
A dynamic bolus phantom for the evaluation of the spatio-temporal resolution of MPI scanners. - In: Journal of magnetism and magnetic materials, ISSN 1873-4766, Bd. 519 (2021), 167446
Magnetic particle imaging (MPI) is a tomographic imaging method to determine the spatial distribution of magnetic nanoparticles (MNP) within a defined volume. To evaluate the spatio-temporal resolution of existing MPI scanners, enabling the consistent comparison of the performance of different scanner setups, we developed dynamic MPI measurement phantoms based on segmented flow. These segmented flow phantoms comprise a defined bolus of ferrofluid tracer material, which can be pumped through a tube system with defined velocities. Using a hydrophobic organic carrier oil, cylindrically shaped boluses of different diameter, length, and flow velocity can be emulated. Moving boluses were imaged by different MPI scanner types and the correlation of spatial resolution und velocity of the bolus was investigated. For all bolus dimension and flow velocity combinations investigated, we observed a decreasing spatial resolution and increasing blurring for increasing bolus velocity and decreasing bolus volume.
https://doi.org/10.1016/j.jmmm.2020.167446
"Hidden" integration of industrial design-tools in E-learning environments. - In: Cross reality and data science in engineering, (2021), S. 437-455
One of the problems that has emerged during the intensive usage of the GOLDi remote lab over the years is the usage of external third-party development tools, necessary for the software- or hardware-oriented design. The installation and setup (e.g. of Atmel Studio or Intel Quartus Prime) on a private PC or laptop is quite complicated, requires some expert knowledge to use these tools safely and effectively, and is not always platform-independent. Above this, these IDE occupy several gigabytes in memory size and sometimes take hours to install. These very powerful tools, of which only a fraction of the offered functionality is needed to solve the given educational task, generate a large overhead. In addition, high school and university students may have only little or no knowledge of microcontrollers, FPGA, compiler linker toolchain, hardware-related languages, etc.That is why we developed a new tool as an integral part of the GOLDi remote lab infrastructure - called WIDE (WEB IDE) - which will be described in this paper. WIDE supports all the design flows with the only requirement of an Internet browser. This means WIDE is running in standalone mode to write and compile code in the specific language or directly inside the Experiment Control Panel (ECP) of the student's Web browser.
Virtual control units in remote labs. - In: Cross reality and data science in engineering, (2021), S. 223-231
Enforced by current results of research in the field of the Internet of Things, working with virtual models has taken a significant step forward. To use such technologies in interactive collaborative and blended learning scenarios, virtual, remote and hybrid online labs has a long tradition in the engineering education process. In virtual or hybrid labs, parts of the experiments take place not on real, but on virtual instruments. In the paper, we present virtual design tools, used in a hybrid online lab that supports a course in designing combinational and sequential logic for control systems. We describe a new integrated hardware-oriented design tool called BEAST (Block-diagram Editing And Simulation Tool) supporting a circuit-oriented design process and works as virtual control unit in a hybrid online lab. We also discuss in which learning scenarios virtual control units have proven themselves.
Preface magnetic carriers conference 2018. - In: Journal of magnetism and magnetic materials, ISSN 1873-4766, Bd. 494 (2020), 165748
https://doi.org/10.1016/j.jmmm.2019.165748