95 research outputs found
Exploring the Effectiveness of Hatha Yoga as a Complementary Treatment for Adolescent Idiopathic Scoliosis: Clinical Effect and Future Research Directions
No full textAdolescent idiopathic scoliosis (AIS) is a common spinal deformity that primarily affects adolescents during the key period of growth and development. While traditional treatment methods often involve bracing or surgery, Hatha yoga, a millennia-old practice rooted in Indian tradition, has emerged as a complementary option for AIS cases. This paper explores the potential benefits of Hatha yoga for adolescents with AIS. It also discusses the limitations of existing research, such as the lack of large-scale randomized controlled trials (RCTs), varying yoga protocols, and challenges in blinding participants and researchers. To address these limitations, I propose future research directions, including conducting large-scale RCTs, long-term follow-up studies, standardized yoga protocols, and assessing safety concerns. I also highlight the need for tailored interventions and comparative effectiveness studies to better understand the potential of Hatha yoga in the holistic treatment of AIS in adolescents
Research of specific field ultra-short text classification based on collaborative filtering algorithm
No full textIn some specific fields, there are a lot of ultra-short texts that need to be categorized. This paper proposes an ultra-short text classification method based on collaborative filtering algorithm aiming at the problems such as short text content, short length, sparse features, and large number of categories in certain fields. First, converting ultra-short text into word frequency vector by doing Chinese word segmentation and calculating word frequency; Secondly, combining relevant data in specific fields, defining the ultra-short texts as users, categories as items, and then constructing a user-item recommendation matrix. Finally, calculating text similarity by using cosine similarity method and obtaining the classification results. The experimental results show that the proposed method can well solve the problem of classification of ultra-short texts in specific fields, and the average accuracy is 9.19% and 3.81% higher than vector space model and topic similarity method respectively
Running CNN efficiently on a FPGA
No full textWith increased demand for AI at the edge, there is a pressing need to adapt
ever more computationally demanding deep learning models for deployment
onto embedded devices. As accelerators for these networks, FPGAs have
become preferred for their energy efficiency and adaptability, but models also
need to be pre-processed before effective FPGA-based hardware accelerators
can be designed. In this project, the author investigates the performance of
Block-Balanced Sparsity, a model compression approach that prunes parameter
matrices in deep learning networks via a structured manner that allows for
efficient FPGA accelerator implementations. By testing this approach across
different pruning strategies, the author found that the fine-tuning strategy led
to the highest model accuracy, gradual pruning allowed for the fastest model
development and learning rate rewinding provided the greatest ease-of-use.Bachelor of Engineering (Computer Engineering
Energy Harvesting PMIC Design for an Extended Power-Range
No full textThis work proposes an energy harvesting DC-DC converter that is able to efficiently process a wide input power range from 25 mW up to 250 mW, to charge a storage device used in Internet-of-Things (IoT) applications. An interleaved boost-converter topology with two inductors is used to divide the high currents over the two branches, in order to reduce the conduction losses in the MOSFET power switches at high input power.A Maximum Power Point Tracking (MPPT) strategy is employed to find the optimal matching between the harvesting sensor, which in this case is a solar cell, and the input of the converter. The Maximum Power Point (MPP) is found by measuring the output current of the converter, decreasing the duty cycle of the switching power MOSFETs, and again measuring the output current to check if it increases. This process is repeated until the output current no longer increases, which is where the MPP will be. The duty-cycle pulses that control the power switches are generated by comparing a reference voltage representing the duty cycle with two sawtooth waves with a phase difference of 180 degrees generated on-chip.Usually, for converters that handle large currents, large power switches are required to keep the on-resistance of the switches minimal such that the conduction losses do not dominate. Since the converter also needs to be able to efficiently convert smaller powers, the size of the power switches is made configurable. This is necessary since otherwise the gate-charge losses will become dominant. The total size of the segmented power switches are controlled digitally by a logic unit that calculates the expected input current using 2-bit representations of the measured output current and the two most significant bits of the 5-bit representation for the duty cycle set by the MPPT.Schematic simulation results show a conversion efficiency of up to 92%. The proposed system is designed and simulated using 180 nm TSMC CMOS technology. The chip covers a total silicon area of 2.20 mm2, with an active area of 1.12 mm2.Electrical Engineerin
[Beitrag zu der untersuchung des pathologischen zustandes der zahne und der Weichen teile der mundhohle bei spastischen kindern]
Get PDFAutor je izvršio sistematske preglede zubi i mekih dijelova usne šupljine jedne grupe psihofizički zaostale djece (uslijed spastične cerebralne dječje klijenuti) i to 18 muških i 22 ženske osobe, ukupno 40. Tom prilikom je ustanovio da je učestalost zubnog karijesa, oboljenja sluznica organa usne šupljine i parodoncija tih osoba bila znatno veća nego ona kontrolne grupe zdrave djece. Ti su nalazi po prevalenciji spomenutih patoloških pojava bili slični ranije dobivenim rezultatima pregleda usne šupljine jedne druge grupe defektne (duševno zaostale) djece.The author checked methodically the teeth and soft parts of oral cavity in spastic children having sequels of cerebral palsy (1 8 boys, 22 girls, totaling to 40 children ) stating the incidency of decayed teeth and pathological alterations in the soft parts of the mouth to be considerably higher than in healthy children. These statings were in prevalency and inten sity of pathological alterations and were similar to the results of checking and controlling teeth and soft parts of the mouth hole in another group of handicaped children, previously examined.Der Autor hat systematische Untersuchungen der Zahne und der weichen Teile der Mundhohle bei einer Gruppe von psychophysisch defekten Kindern (in folge spastischer zerebraler Kinderlahmung)
und zwar bei 18 Knaben, 22 Madchen, insgesamt 40 Kindern durchge fuhrt. Dabei hat er festgestellt, dass Erscheinungen von Zahnkaries, Erkrankungen der Schleimhaut des Mundhole und des Parodontium bei solchen Kindern bedeutend haufiger vorkommen als bei untersuchten gesunden Kindern. Diese Befunde waren in der Vorherrsch aft der genannten pathologischen Erscheinungen ahnlich den fruher erlangten Resultaten bei Untersuchungen der Mundhohle bei einer anderen Gruppe von defekten (geistig zuruckgebliebenen) Kindern
Soft UV nanoimprint lithography: concept, development, and fabrication of nanostructures with tunable feature sizes at constant pitch
Get PDFThe thrilling development of the nanoimprint lithography (NIL) over the past twenty years has made the large area nanopatterning from an obstacle to the bridge between imagination and reality. This thesis addresses the establishment of NIL technique at the Institute of Micro- and Nanotechnologies (IMN) of Technische Universität Ilmenau and the development of the soft UV-NIL. The fabrications of periodic nanostructures with tunable feature sizes in large area and at low cost as well as its additional applications are investigated. A center-to-edge imprint scheme in ambient atmosphere using soft stamps is employed throughout this work. The advantages of soft UV-NIL have been identified all along and have drawn tremendous research attention. To clarify the potential issues for imprint in ambient atmosphere, a type of bi-layer soft stamp with a soft Polydimethylsiloxane (PDMS) back carrier and a thin feature layer employing various material is configured. The bi-layer soft stamps such as PDMS/PDMS, PDMS/organic solvent diluted PDMS, PDMS/X-PDMS, PDMS/vvsPDMS, PDMS/UV-PDMS, are comparatively characterized. High stamp reusability and imprint uniformity of the PDMS/PDMS, PDMS/toluene-diluted PDMS and PDMS/UV-PDMS stamps have been validated at ambient conditions for at least twenty consecutive imprints by using a single stamp. This demonstrates the suitability of NIL for batch fabrication of nanoimprinted structures on wafer scale. Based on the solid imprint performance of the ambient center-to-edge scheme, a process chain to fabricate nanostructures with tunable feature sizes at constant pitch (NanoTuFe) is designed and realized in this thesis. A master with positive or negative nanopatterns is prepared to initialize the fabrication process. An intermediate template featuring sloping sidewalls is introduced to bridge the original master and the final patterns. Taking advantage of the sloping sidewalls of the patterns on the intermediate template, the etch-mask on the final substrate can be opened to discrete dimensions simply by varying the etching durations. Numerous new NIL templates and substrates with diverse layouts, profiles and aspect ratios have been achieved with tunable feature sizes. - New NIL templates of circular nanopillars with diameters of 150/200/250/350 nm and aspect ratio below 2 are generated for NIL templates based on a master featuring 450 nm circular nanopillars. - Additional NIL templates of square nanocavities with 130/160/190/220 nm feature sizes are fabricated from a master patterned by circular nanocavities of 350 nm in diameter - By introducing a bi-layer lift-off process, the image tone, layout and feature sizes can be adjusted. NIL templates featuring square nanopillars with dimensions of 120/200/320 nm are achieved based on a master of circular nanoholes with 350 nm in diameter. Furthermore, diverse top-down or bottom-up surface patterning techniques can be integrated to the NanoTuFe methodology. The versatility using NanoTuFe to fabricate high aspect ratio nanostructures (feature sizes from 70 nm to 400 nm, etch depth of 4-7 mm) for high emissive silicon surfaces in both circular and square layouts is demonstrated. The concept of using NIL for curved surface (convex-plano refractive lens) has been proved as well. Replicating nanostructures is in general cost-intensive, especially for large patterning area and with high resolution. The NanoTuFe approach enables a fast realization of large area nanopatterns with tunable dimensions at low cost, which is one of the essential contributions of this work.Die vorliegende Dissertation beschäftigt sich mit der Weiterentwicklung der UV Nanoimprint-Lithographie (NIL) mit weichen Stempeln am Institut für Mikro- und Nanotechnologien (IMN) der Technischen Universität Ilmenau. Die Herstellung von periodischen Nanostrukturen mit einstellbaren Strukturgrößen auf großer Fläche und zu niedrigen Kosten sowie deren weiterführende Anwendungen werden untersucht. Allen hier vorgestellten Arbeiten liegt die UV-NIL mit weichen Stempeln bei Umgebungsatmosphäre zugrunde, bei welcher der eigentliche Abformvorgang von der Stempelmitte zum Rand fortschreitet (engl. center-to-edge). Die Vorteile von UV-NIL mit weichen Stempeln sind lange bekannt und haben eine hohe Aufmerksamkeit auf sich gezogen. Zweischichtige weiche Stempel mit einem weichen Polydimethylsiloxane (PDMS)-Träger und einer dünnen Strukturschicht, die aus unterschiedlichen und festeren Materialien bestehen kann, werden konfiguriert. Die zweischichtigen weichen Stempel wie PDMS/PDMS, PDMS/verdünntes PDMS, PDMS/X-PDMS, PDMS/vvsPDMS, PDMS/UV-PDMS, werden miteinander verglichen. Hohe Gleichmäßigkeit der Abformungen der PDMS/PDMS, PDMS/Toluen-verdünnten PDMS und PDMS/UV-PDMS Stempel wurden bei Umgebungsbedingungen für mindestens zwanzig aufeinanderfolgende Prozesse in Bezug auf die Wiederholbarkeit validiert. Basierend auf dem beschriebenen Abform-Regime, wird in dieser Arbeit eine Prozesskette zur Herstellung von Nanostrukturen mit einstellbaren Strukturgrößen und -formen (engl. NanoTuFe) realisiert. Eine Zwischenmaske mit positiven oder negativen Strukturen mit geneigten Seitenwänden wird eingeführt, um die Brücke zwischen ursprünglichem Master und den herzustellenden Strukturen zu schlagen. Unter Ausnutzung der geneigten Seitenwände der Strukturen auf der Zwischenschablone kann die Ätzmaske auf dem finalen Substrat durch Variation der Ätzdauer in sehr variable Dimensionen überführt werden. Zahlreiche NILMasken und -Substrate, die einstellbare Strukturgrößen verschiedener Layouts, Profile und Aspektverhältnisse aufweisen, wurden untersucht. - Es werden Neue NIL-Masken von kreisförmigen Nanosäulen mit Durchmessern von 150/200/250/350 nm und einem Aspektverhältnis unter 2 basierend auf einem Master mit 450 nm langen kreisförmigen Nanosäulen erzeugt. – Zusätzliche NIL-Masken mit quadratischen Nanokavitäten mit Strukturgrößen von 130/160/190/220 nm werden aus einem Master mit kreisförmigen Nanokavitäten von 350 nm Durchmesser hergestellt. - Durch die Einführung eines doppelschichtigen Lift-Off-Prozesses kann eine positive oder negative Abbildung der Strukturen erfolgen sowie die Layout- und Feature-Größe angepasst werden. NIL-Masken mit quadratischen Nanosäulen mit Abmessungen von 120/200/320 nm werden anhand eines Masters aus kreisförmigen Nanolöchern mit 350 nm Durchmesser demonstriert. Darüber hinaus können verschiedene Top-Down oder Bottom-Up Strukturierungstechniken in die NanoTuFe - Methode integriert werden. Die Vielseitigkeit von NanoTuFe zur Herstellung von Nanostrukturen mit großem Aspektverhältnis (Strukturgrößen von 70 nm bis 400 nm, Ätztiefe von 4-7 mm) für hoch emittierende Siliziumoberflächen in kreisförmigen und quadratischen Layouts wird demonstriert. Das Konzept der Verwendung von NIL auf gekrümmte Oberflächen (konvex-plane refraktive Linse) wurde ebenfalls bewiesen. Die Herstellung und Replikation von Nanostrukturen ist im Allgemeinen kostenintensiv, insbesondere für große Flächen und bei hoher Auflösung. Die NanoTuFe-Methode ermöglicht die schnelle Realisierung großflächiger Nanostrukturen mit einstellbaren Abmessungen und Formen (Kreise, Rechtecke, Säulen, Kegel, Pyramiden) bei potenziell geringen Kosten, was einer der wesentlichen Beiträge dieser Arbeit ist
Bildung und Stabilität von Phasen aus Li, Sn, Si und C
Get PDFPhasenbildungsprozesse und Phasenstabilitäten wurden im binären System Li-C sowie in den ternären Systemen Li-C-Si, Li-C-Sn und Li-Si-Sn untersucht. Im binären Li-C-System lagen die Schwerpunkte der Untersuchungen auf der Charakterisierung der batterierelevanten Li-Graphit-Interkalationsphasen (Li-GICs) sowie auf der Untersuchung der thermodynamischen Stabilität der in diesem System bekannten Phasen LiC6 und Li2C2. Mit Hilfe von Wärmebehandlungsexperimenten und dynamischer Differenzkalorimetrie konnte gezeigt werden, dass die Phase Li2C2 thermodynamisch stabil ist, wohingegen es sich bei der Li-Graphit-Interkalationsphase LiC6 um eine metastabile Phase handelt. Erstmals wurden Gefügeuntersuchungen in Verbindung mit Röntgendiffraktometrie eingesetzt, um Li-GICs mit verschiedenen Li-Gehalten zu charakterisieren. Es wurde gezeigt, dass die Bildung der Li-GICs während des Ein- und Ausbaus von Li nicht in allen Aspekten symmetrisch verläuft. Die Untersuchungen zeigten außerdem, dass die Phase LiC12 einen variablen Li-Gehalt aufweisen kann und somit nicht stöchiometrisch ist. Die bislang komplett unbeschriebenen ternären Systeme Li-Si-C, Li-Sn-C und Li-Si-Sn wurden mittels Wärmebehandlungsexperimenten und dynamischer Differenzkalorimetrie hinsichtlich ihrer Phasengleichgewichte untersucht. Hierfür wurden gezielt bestimmte Zusammensetzungen ausgewählt, welche mittels binärer Phasen aus den Randsystemen präzise eingestellt wurden. Die bei der Untersuchung der ternären Systeme detektierten Phasen weichen deutlich von denen ab, die durch Extrapolation aus den binären Randsystemen vorhergesagt werden. Im System Li-Si-C wurde ein Dreiphasengleichgewicht, welches die Phasen Li7Si3, Li2C2 und SiC enthält, eindeutig identifiziert. Für eine abschließende Aussage über die im thermodynamischen Gleichgewicht vorliegenden Phasen in den Systemen Li-Sn-C und Li-Sn-Si bedarf es einer weiteren, detaillierten Untersuchung. Ternäre Phasen wurden in keinem der betrachteten Systeme beobachtet
Interactive Fabrication: Exploration of Human-machine Collaboration Methods for AR Assisted Robotic Fabrication
Get PDFLAUREA MAGISTRALENegli ultimi anni, la ricerca nella progettazione computazionale e nella fabbricazione robotica in architettura, ingegneria e costruzione (AEC) ha fatto notevoli progressi nell'automazione dei processi di costruzione. Tuttavia, sono state fatte poche ricerche su come sfruttare i processi umani-in-the-loop per scenari di fabbricazione robotica su larga scala. In tali processi, umani e robot si supportano a vicenda in operazioni di fabbricazione che nessuno di loro poteva gestire da solo, portando a nuove opportunità per il dominio AEC.
Allo stesso tempo, a causa degli aggiornamenti dei dispositivi e della maturità tecnologica, l'uso della tecnologia di realtà aumentata (AR) è aumentato. Per la progettazione architettonica, in particolare per i progetti di fabbricazione digitale, è iniziata una maggiore ricerca per integrare metodi AR per ottenere la visualizzazione nel processo. Recentemente, ci sono stati esperimenti e ricerche pratiche sulla fabbricazione digitale di edifici basati su AR, che è un metodo di fabbricazione interattivo. La natura intuitiva della AR consente a esseri umani e robot di interagire direttamente nello spazio reale utilizzando movimenti degli arti, che è un approccio di programmazione robot più diretto e a soglia inferiore rispetto al controllo dei robot tramite codice su un computer. Pertanto, questo è un ottimo metodo per incorporare la creatività e l'intelligenza umana nella produzione lineare di robot su larga scala.
Sulla base dei problemi di cui sopra, questo articolo continua ad esplorare il metodo della costruzione di robot assistiti AR. In primo luogo, è stato introdotto lo sfondo dell'argomento, e poi è stato completato un esperimento di posizionamento di blocchi di legno utilizzando robot di insegnamento dipendente, strumenti computazionali generali e strumenti di programmazione AR, per confrontare le differenze tra questo nuovo metodo e metodi tradizionali. Allo stesso tempo, le nuove possibilità che possono essere fornite nel processo di costruzione di robot assistiti AR sono state riassunte attraverso esperimenti, come interfacce di controllo utente virtuali, controllo gestuale dei percorsi di movimento del robot, anteprima dei processi di lavoro del robot e così via.
Sulla base di questi metodi, sono stati completati due progetti di costruzione interattivi. Il primo viene utilizzato per lucidare motivi decorativi metallici su facciate e pareti, e il secondo è una costruzione speciale di mattoni in stile basato su mattoni inclinati di robot mobili. Nel primo progetto, la tecnologia AR è stata utilizzata per completare attività come disegnare percorsi di lucidatura, impostare altri parametri robot come la velocità e visualizzare in anteprima muri virtuali; Nel secondo progetto, la tecnologia AR è stata utilizzata per completare attività come il posizionamento del robot mobile, la selezione dei mattoni per il posizionamento e l'anteprima del processo di movimento del robot.
In questi progetti, questo articolo esplora la possibilità di applicare la tecnologia AR alla costruzione di robot, consentendo agli esseri umani di partecipare maggiormente alla costruzione di robot lineari per rendere il sistema di costruzione più adattabile e creativo. L'applicazione della tecnologia AR nel progetto ha anche verificato il suo enorme potenziale nel campo della fabbricazione digitale. In futuro, con l'ulteriore sviluppo della tecnologia, la fabbricazione robotica assistita da AR dovrebbe avere un impatto significativo sul settore AEC.In recent years, research in computational design and robotic fabrication in architecture, engineering, and construction (AEC) has made remarkable advances in automating construction processes. However, little research has been done on how to leverage human-in-the-loop processes for large-scale robotic fabrication scenarios. In such processes, humans and robots support each other in fabrication operations that neither of them could handle alone, leading to new opportunities for the AEC domain.
At the same time, due to device updates and technological maturity, the use of augmented reality (AR) technology has increased. For architectural design, especially digital fabrication projects, more research has begun to integrate AR methods to achieve visualization in the process. Recently, there have been experiments and practical research on digital fabrication of buildings based on AR, which is an interactive fabrication method. The intuitive nature of AR allows humans and robots to directly interact in real space using limb movements, which is a more direct and lower threshold robot programming approach compared to controlling robots through code on a computer. Therefore, this is a very good method to incorporate human creativity and intelligence into linear large-scale robot manufacturing.
Based on the above issues, this paper continues to explore the method of AR assisted robot construction. Firstly, the background of the topic was introduced, and then a wooden block placement experiment was completed using robot teaching pendent, general computational tools, and AR programming tools, to compare the differences between this new method and traditional methods. At the same time, new possibilities that can be provided in the AR assisted robot construction process were summarized through experiments, such as virtual user control interfaces, gesture control of robot movement paths, preview of robot work processes, and so on.
Based on the above methods, two interactive construction projects were completed. The first is used for polishing metal decorative patterns on facades and walls, and the second is a special style brick wall construction based on inclined bricks of mobile robots. In the first project, AR technology was used to complete tasks such as drawing polishing paths, setting other robot parameters such as speed, and previewing virtual walls; In the second project, AR technology was used to complete tasks such as positioning the mobile robot, selecting bricks for placement, and previewing the robot's motion process.
In these projects, this paper explores the possibility of applying AR technology to robot construction, allowing humans to participate more in linear robot construction to make the construction system more adaptable and creative. The application of AR technology in the project has also verified its enormous potential in the field of digital fabrication. In the future, with the further development of technology, AR assisted robotic fabrication is expected to have a significant impact on the AEC industry
Reduction kinetics of Wüstite scale on pure iron and steel sheets in Ar and H<sub>2</sub> gas mixture
No full textA dense and closed Wüstite scale is formed on pure iron and Mn alloyed steel after oxidation in Ar + 33 vol pct CO2 + 17 vol pct CO gas mixture. Reducing the Wüstite scale in Ar + H2 gas mixture forms a dense and uniform iron layer on top of the remaining Wüstite scale, which separates the unreduced scale from the gas mixture. The reduction of Wüstite is controlled by the bulk diffusion of dissolved oxygen in the formed iron layer and follows parabolic growth rate law. The reduction kinetics of Wüstite formed on pure iron and on Mn alloyed steel are the same. The parabolic rate constant of Wüstite reduction obeys an Arrhenius relation with an activation energy of 104 kJ/mol if the formed iron layer is in the ferrite phase. However, at 1223 K (950 °C) the parabolic rate constant of Wüstite reduction drops due to the phase transformation of the iron layer from ferrite to austenite. The effect of oxygen partial pressure on the parabolic rate constant of Wüstite reduction is negligible when reducing in a gas mixture with a dew point below 283 K (10 °C). During oxidation of the Mn alloyed steel, Mn is dissolved in the Wüstite scale. Subsequently, during reduction of the Wüstite layer, Mn diffuses into the unreduced Wüstite. Ultimately, an oxide-free iron layer is obtained at the surface of the Mn alloyed steel, which is beneficial for coating application.</p
- …
