Ferramenta computacional para identificação de micro-organismos com base em assinaturas genômicas

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Desenvolvimento de um sistema de telemetria para monitoramento térmico em casas de vegetação

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Desenvolvimento de dispositivos computacionais para registro de dados pluviométricos e análise operacional de pluviômetros basculantes

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Sistema informatizado para dispositivos móveis de coleta dos dados referentes à cesta básica do município de Botucatu/SP

The collection of prices for basic goods supply is very important for the population, based on the collection and processing of these data the CLI (Cost Living Index) is calculated among others, helping consumers to shop more rationally and with a clearer view of each product impact of each product on their household budget, not only food, but also cleaning products and personal hygiene ones. Nowadays, the project of collection of prices for basic goods supply is conducted weekly in Botucatu - SP through a spreadsheet. The aim of this work was to develop a software which utilized mobile devices in the data collection and storage phase, concerning the basic goods supply in Botucatu -SP. This was created in order to eliminate the need of taking notes in paper spreadsheets, increasing efficiency and accelerating the data processing. This work utilized the world of mobile technology and development tools, through the platform".NET" - Compact Framework and programming language Visual Basic".NET" was used in the handheld phase, enabling to develop a system using techniques of object oriented programming, with higher speed and reliability in the codes writing. A HP Pavilion dv3 personal computer and an Eten glofish x500+ handheld computer were used. At the end of the software development, collection, data storing and processing in a report, the phase of in loco paper spreadsheets were eliminated and it was possible to verify that the whole process was faster, more consistent, safer, more efficient and the data were more available.

Ultrasound Transient Shear Wave Elasticity Imaging for Tendon Tissue

Degeneration of tendon tissue is a common cause of tendon dysfunction with the symptoms of repeated episodes of pain and palpable increase of tendon thickness. Tendon mechanical properties are directly related to its physiological composition and the structural organization of the interior collagen fibers which could be altered by tendon degeneration due to overuse or injury. Thus, measuring mechanical properties of tendon tissue may represent a quantitative measurement of pain, reduced function, and tissue health. Ultrasound elasticity imaging has been developed in the last two decades and has proved to be a promising tool for tissue elasticity imaging. To date, however, well established protocols of tendinopathy elasticity imaging for diagnosing tendon degeneration in early stages or late stages do not exist. This thesis describes the re-creation of one dynamic ultrasound elasticity imaging method and the development of an ultrasound transient shear wave elasticity imaging platform for tendon and other musculoskeletal tissue imaging. An experimental mechanical stage with proper supporting systems and accurate translating stages was designed and made. A variety of high-quality tissue-mimicking phantoms were made to simulate homogeneous and heterogeneous soft tissues as well as tendon tissues. A series of data acquisition and data processing programs were developed to collect the displacement data from the phantom and calculate the shear modulus and Young’s modulus of the target. The imaging platform was found to be capable of conducting comparative measurements of the elastic parameters of the phantoms and quantitatively mapping elasticity onto ultrasound B-Mode images. This suggests the system has great potential for not only benefiting individuals with tendinopathy with an earlier detection, intervention and better rehabilitation, but also for providing a medical tool for quantification of musculoskeletal tissue dysfunction in other regions of the body such as the shoulder, elbow and knee.

A palaeobiologist's guide to 'virtual' micro-CT preparation

This paper provides a brief but comprehensive guide to creating, preparing and dissecting a 'virtual' fossil, using a worked example to demonstrate some standard data processing techniques. Computed tomography (CT) is a 3D imaging modality for producing 'virtual' models of an object on a computer. In the last decade, CT technology has greatly improved, allowing bigger and denser objects to be scanned increasingly rapidly. The technique has now reached a stage where systems can facilitate large-scale, non-destructive comparative studies of extinct fossils and their living relatives. Consequently the main limiting factor in CT-based analyses is no longer scanning, but the hurdles of data processing (see disclaimer). The latter comprises the techniques required to convert a 3D CT volume (stack of digital slices) into a virtual image of the fossil that can be prepared (separated) from the matrix and 'dissected' into its anatomical parts. This technique can be applied to specimens or part of specimens embedded in the rock matrix that until now have been otherwise impossible to visualise. This paper presents a suggested workflow explaining the steps required, using as example a fossil tooth of Sphenacanthus hybodoides (Egerton), a shark from the Late Carboniferous of England. The original NHMUK copyrighted CT slice stack can be downloaded for practice of the described techniques, which include segmentation, rendering, movie animation, stereo-anaglyphy, data storage and dissemination. Fragile, rare specimens and type materials in university and museum collections can therefore be virtually processed for a variety of purposes, including virtual loans, website illustrations, publications and digital collections. Micro-CT and other 3D imaging techniques are increasingly utilized to facilitate data sharing among scientists and on education and outreach projects. Hence there is the potential to usher in a new era of global scientific collaboration and public communication using specimens in museum collections.

Underwater mapping using sonar

[ES]El proyecto contiene módulos de simulación, procesado de datos, mapeo y localización, desarrollados en C++ utilizando ROS (Robot Operating System) y PCL (Point Cloud Library). Ha sido desarrollado bajo el proyecto de robótica submarina AVORA.Se han caracterizado el vehículo y el sensor, y se han analizado diferentes tecnologías de sensores y mapeo. Los datos pasan por tres etapas: Conversión a nube de puntos, filtrado por umbral, eliminación de puntos espureos y, opcionalmente, detección de formas. Estos datos son utilizados para construir un mapa de superficie multinivel. La otra herramienta desarrollada es un algoritmo de Punto más Cercano Iterativo (ICP) modificado, que tiene en cuenta el modo de funcionamiento del sonar de imagen utilizado.

Sistemi riconfigurabili a basso consumo per applicazioni di monitoraggio distribuito

The term Ambient Intelligence (AmI) refers to a vision on the future of the information society where smart, electronic environment are sensitive and responsive to the presence of people and their activities (Context awareness). In an ambient intelligence world, devices work in concert to support people in carrying out their everyday life activities, tasks and rituals in an easy, natural way using information and intelligence that is hidden in the network connecting these devices. This promotes the creation of pervasive environments improving the quality of life of the occupants and enhancing the human experience. AmI stems from the convergence of three key technologies: ubiquitous computing, ubiquitous communication and natural interfaces. Ambient intelligent systems are heterogeneous and require an excellent cooperation between several hardware/software technologies and disciplines, including signal processing, networking and protocols, embedded systems, information management, and distributed algorithms. Since a large amount of fixed and mobile sensors embedded is deployed into the environment, the Wireless Sensor Networks is one of the most relevant enabling technologies for AmI. WSN are complex systems made up of a number of sensor nodes which can be deployed in a target area to sense physical phenomena and communicate with other nodes and base stations. These simple devices typically embed a low power computational unit (microcontrollers, FPGAs etc.), a wireless communication unit, one or more sensors and a some form of energy supply (either batteries or energy scavenger modules). WNS promises of revolutionizing the interactions between the real physical worlds and human beings. Low-cost, low-computational power, low energy consumption and small size are characteristics that must be taken into consideration when designing and dealing with WSNs. To fully exploit the potential of distributed sensing approaches, a set of challengesmust be addressed. Sensor nodes are inherently resource-constrained systems with very low power consumption and small size requirements which enables than to reduce the interference on the physical phenomena sensed and to allow easy and low-cost deployment. They have limited processing speed,storage capacity and communication bandwidth that must be efficiently used to increase the degree of local ”understanding” of the observed phenomena. A particular case of sensor nodes are video sensors. This topic holds strong interest for a wide range of contexts such as military, security, robotics and most recently consumer applications. Vision sensors are extremely effective for medium to long-range sensing because vision provides rich information to human operators. However, image sensors generate a huge amount of data, whichmust be heavily processed before it is transmitted due to the scarce bandwidth capability of radio interfaces. In particular, in video-surveillance, it has been shown that source-side compression is mandatory due to limited bandwidth and delay constraints. Moreover, there is an ample opportunity for performing higher-level processing functions, such as object recognition that has the potential to drastically reduce the required bandwidth (e.g. by transmitting compressed images only when something ‘interesting‘ is detected). The energy cost of image processing must however be carefully minimized. Imaging could play and plays an important role in sensing devices for ambient intelligence. Computer vision can for instance be used for recognising persons and objects and recognising behaviour such as illness and rioting. Having a wireless camera as a camera mote opens the way for distributed scene analysis. More eyes see more than one and a camera system that can observe a scene from multiple directions would be able to overcome occlusion problems and could describe objects in their true 3D appearance. In real-time, these approaches are a recently opened field of research. In this thesis we pay attention to the realities of hardware/software technologies and the design needed to realize systems for distributed monitoring, attempting to propose solutions on open issues and filling the gap between AmI scenarios and hardware reality. The physical implementation of an individual wireless node is constrained by three important metrics which are outlined below. Despite that the design of the sensor network and its sensor nodes is strictly application dependent, a number of constraints should almost always be considered. Among them: • Small form factor to reduce nodes intrusiveness. • Low power consumption to reduce battery size and to extend nodes lifetime. • Low cost for a widespread diffusion. These limitations typically result in the adoption of low power, low cost devices such as low powermicrocontrollers with few kilobytes of RAMand tenth of kilobytes of program memory with whomonly simple data processing algorithms can be implemented. However the overall computational power of the WNS can be very large since the network presents a high degree of parallelism that can be exploited through the adoption of ad-hoc techniques. Furthermore through the fusion of information from the dense mesh of sensors even complex phenomena can be monitored. In this dissertation we present our results in building several AmI applications suitable for a WSN implementation. The work can be divided into two main areas:Low Power Video Sensor Node and Video Processing Alghoritm and Multimodal Surveillance . Low Power Video Sensor Nodes and Video Processing Alghoritms In comparison to scalar sensors, such as temperature, pressure, humidity, velocity, and acceleration sensors, vision sensors generate much higher bandwidth data due to the two-dimensional nature of their pixel array. We have tackled all the constraints listed above and have proposed solutions to overcome the current WSNlimits for Video sensor node. We have designed and developed wireless video sensor nodes focusing on the small size and the flexibility of reuse in different applications. The video nodes target a different design point: the portability (on-board power supply, wireless communication), a scanty power budget (500mW),while still providing a prominent level of intelligence, namely sophisticated classification algorithmand high level of reconfigurability. We developed two different video sensor node: The device architecture of the first one is based on a low-cost low-power FPGA+microcontroller system-on-chip. The second one is based on ARM9 processor. Both systems designed within the above mentioned power envelope could operate in a continuous fashion with Li-Polymer battery pack and solar panel. Novel low power low cost video sensor nodes which, in contrast to sensors that just watch the world, are capable of comprehending the perceived information in order to interpret it locally, are presented. Featuring such intelligence, these nodes would be able to cope with such tasks as recognition of unattended bags in airports, persons carrying potentially dangerous objects, etc.,which normally require a human operator. Vision algorithms for object detection, acquisition like human detection with Support Vector Machine (SVM) classification and abandoned/removed object detection are implemented, described and illustrated on real world data. Multimodal surveillance: In several setup the use of wired video cameras may not be possible. For this reason building an energy efficient wireless vision network for monitoring and surveillance is one of the major efforts in the sensor network community. Energy efficiency for wireless smart camera networks is one of the major efforts in distributed monitoring and surveillance community. For this reason, building an energy efficient wireless vision network for monitoring and surveillance is one of the major efforts in the sensor network community. The Pyroelectric Infra-Red (PIR) sensors have been used to extend the lifetime of a solar-powered video sensor node by providing an energy level dependent trigger to the video camera and the wireless module. Such approach has shown to be able to extend node lifetime and possibly result in continuous operation of the node.Being low-cost, passive (thus low-power) and presenting a limited form factor, PIR sensors are well suited for WSN applications. Moreover techniques to have aggressive power management policies are essential for achieving long-termoperating on standalone distributed cameras needed to improve the power consumption. We have used an adaptive controller like Model Predictive Control (MPC) to help the system to improve the performances outperforming naive power management policies.

Calcolo della posizione di reti di stazioni permanenti

Con il trascorrere del tempo, le reti di stazioni permanenti GNSS (Global Navigation Satellite System) divengono sempre più un valido supporto alle tecniche di rilevamento satellitare. Esse sono al tempo stesso un’efficace materializzazione del sistema di riferimento e un utile ausilio ad applicazioni di rilevamento topografico e di monitoraggio per il controllo di deformazioni. Alle ormai classiche applicazioni statiche in post-processamento, si affiancano le misure in tempo reale sempre più utilizzate e richieste dall’utenza professionale. In tutti i casi risulta molto importante la determinazione di coordinate precise per le stazioni permanenti, al punto che si è deciso di effettuarla tramite differenti ambienti di calcolo. Sono stati confrontati il Bernese, il Gamit (che condividono l’approccio differenziato) e il Gipsy (che utilizza l’approccio indifferenziato). L’uso di tre software ha reso indispensabile l’individuazione di una strategia di calcolo comune in grado di garantire che, i dati ancillari e i parametri fisici adottati, non costituiscano fonte di diversificazione tra le soluzioni ottenute. L’analisi di reti di dimensioni nazionali oppure di reti locali per lunghi intervalli di tempo, comporta il processamento di migliaia se non decine di migliaia di file; a ciò si aggiunge che, talora a causa di banali errori, oppure al fine di elaborare test scientifici, spesso risulta necessario reiterare le elaborazioni. Molte risorse sono quindi state investite nella messa a punto di procedure automatiche finalizzate, da un lato alla preparazione degli archivi e dall’altro all’analisi dei risultati e al loro confronto qualora si sia in possesso di più soluzioni. Dette procedure sono state sviluppate elaborando i dataset più significativi messi a disposizione del DISTART (Dipartimento di Ingegneria delle Strutture, dei Trasporti, delle Acque, del Rilevamento del Territorio - Università di Bologna). E’ stato così possibile, al tempo stesso, calcolare la posizione delle stazioni permanenti di alcune importanti reti locali e nazionali e confrontare taluni fra i più importanti codici scientifici che assolvono a tale funzione. Per quanto attiene il confronto fra i diversi software si è verificato che: • le soluzioni ottenute dal Bernese e da Gamit (i due software differenziati) sono sempre in perfetto accordo; • le soluzioni Gipsy (che utilizza il metodo indifferenziato) risultano, quasi sempre, leggermente più disperse rispetto a quelle degli altri software e mostrano talvolta delle apprezzabili differenze numeriche rispetto alle altre soluzioni, soprattutto per quanto attiene la coordinata Est; le differenze sono però contenute in pochi millimetri e le rette che descrivono i trend sono comunque praticamente parallele a quelle degli altri due codici; • il citato bias in Est tra Gipsy e le soluzioni differenziate, è più evidente in presenza di determinate combinazioni Antenna/Radome e sembra essere legato all’uso delle calibrazioni assolute da parte dei diversi software. E’ necessario altresì considerare che Gipsy è sensibilmente più veloce dei codici differenziati e soprattutto che, con la procedura indifferenziata, il file di ciascuna stazione di ciascun giorno, viene elaborato indipendentemente dagli altri, con evidente maggior elasticità di gestione: se si individua un errore strumentale su di una singola stazione o se si decide di aggiungere o togliere una stazione dalla rete, non risulta necessario il ricalcolo dell’intera rete. Insieme alle altre reti è stato possibile analizzare la Rete Dinamica Nazionale (RDN), non solo i 28 giorni che hanno dato luogo alla sua prima definizione, bensì anche ulteriori quattro intervalli temporali di 28 giorni, intercalati di sei mesi e che coprono quindi un intervallo temporale complessivo pari a due anni. Si è così potuto verificare che la RDN può essere utilizzata per l’inserimento in ITRF05 (International Terrestrial Reference Frame) di una qualsiasi rete regionale italiana nonostante l’intervallo temporale ancora limitato. Da un lato sono state stimate le velocità ITRF (puramente indicative e non ufficiali) delle stazioni RDN e, dall’altro, è stata effettuata una prova di inquadramento di una rete regionale in ITRF, tramite RDN, e si è verificato che non si hanno differenze apprezzabili rispetto all’inquadramento in ITRF, tramite un congruo numero di stazioni IGS/EUREF (International GNSS Service / European REference Frame, SubCommission for Europe dello International Association of Geodesy).

Verkehrsmonitoring mit Hilfe von Radarfernerkundung

Für die Zukunft wird eine Zunahme an Verkehr prognostiziert, gleichzeitig herrscht ein Mangel an Raum und finanziellen Mitteln, um weitere Straßen zu bauen. Daher müssen die vorhandenen Kapazitäten durch eine bessere Verkehrssteuerung sinnvoller genutzt werden, z.B. durch Verkehrsleitsysteme. Dafür werden räumlich aufgelöste, d.h. den Verkehr in seiner flächenhaften Verteilung wiedergebende Daten benötigt, die jedoch fehlen. Bisher konnten Verkehrsdaten nur dort erhoben werden, wo sich örtlich feste Meßeinrichtungen befinden, jedoch können damit die fehlenden Daten nicht erhoben werden. Mit Fernerkundungssystemen ergibt sich die Möglichkeit, diese Daten flächendeckend mit einem Blick von oben zu erfassen. Nach jahrzehntelangen Erfahrungen mit Fernerkundungsmethoden zur Erfassung und Untersuchung der verschiedensten Phänomene auf der Erdoberfläche wird nun diese Methodik im Rahmen eines Pilotprojektes auf den Themenbereich Verkehr angewendet. Seit Ende der 1990er Jahre wurde mit flugzeuggetragenen optischen und Infrarot-Aufnahmesystemen Verkehr beobachtet. Doch bei schlechten Wetterbedingungen und insbesondere bei Bewölkung, sind keine brauchbaren Aufnahmen möglich. Mit einem abbildenden Radarverfahren werden Daten unabhängig von Wetter- und Tageslichtbedingungen oder Bewölkung erhoben. Im Rahmen dieser Arbeit wird untersucht, inwieweit mit Hilfe von flugzeuggetragenem synthetischem Apertur Radar (SAR) Verkehrsdaten aufgenommen, verarbeitet und sinnvoll angewendet werden können. Nicht nur wird die neue Technik der Along-Track Interferometrie (ATI) und die Prozessierung und Verarbeitung der aufgenommenen Verkehrsdaten ausführlich dargelegt, es wird darüberhinaus ein mit dieser Methodik erstellter Datensatz mit einer Verkehrssimulation verglichen und bewertet. Abschließend wird ein Ausblick auf zukünftige Entwicklungen der Radarfernerkundung zur Verkehrsdatenerfassung gegeben.

Analyse von Middleware-Technologien bzgl. Performanz und ihrer Eignung für lose und eng gekoppelte verteilte Anwendungen

Die vorliegende Dissertation analysiert die Middleware- Technologien CORBA (Common Object Request Broker Architecture), COM/DCOM (Component Object Model/Distributed Component Object Model), J2EE (Java-2-Enterprise Edition) und Web Services (inklusive .NET) auf ihre Eignung bzgl. eng und lose gekoppelten verteilten Anwendungen. Zusätzlich werden primär für CORBA die dynamischen CORBA-Komponenten DII (Dynamic Invocation Interface), IFR (Interface Repository) und die generischen Datentypen Any und DynAny (dynamisches Any) im Detail untersucht. Ziel ist es, a. konkrete Aussagen über diese Komponenten zu erzielen, und festzustellen, in welchem Umfeld diese generischen Ansätze ihre Berechtigung finden. b. das zeitliche Verhalten der dynamischen Komponenten bzgl. der Informationsgewinnung über die unbekannten Objekte zu analysieren. c. das zeitliche Verhalten der dynamischen Komponenten bzgl. ihrer Kommunikation zu messen. d. das zeitliche Verhalten bzgl. der Erzeugung von generischen Datentypen und das Einstellen von Daten zu messen und zu analysieren. e. das zeitliche Verhalten bzgl. des Erstellens von unbekannten, d. h. nicht in IDL beschriebenen Datentypen zur Laufzeit zu messen und zu analysieren. f. die Vorzüge/Nachteile der dynamischen Komponenten aufzuzeigen, ihre Einsatzgebiete zu definieren und mit anderen Technologien wie COM/DCOM, J2EE und den Web Services bzgl. ihrer Möglichkeiten zu vergleichen. g. Aussagen bzgl. enger und loser Koppelung zu tätigen. CORBA wird als standardisierte und vollständige Verteilungsplattform ausgewählt, um die o. a. Problemstellungen zu untersuchen. Bzgl. seines dynamischen Verhaltens, das zum Zeitpunkt dieser Ausarbeitung noch nicht oder nur unzureichend untersucht wurde, sind CORBA und die Web Services richtungsweisend bzgl. a. Arbeiten mit unbekannten Objekten. Dies kann durchaus Implikationen bzgl. der Entwicklung intelligenter Softwareagenten haben. b. der Integration von Legacy-Applikationen. c. der Möglichkeiten im Zusammenhang mit B2B (Business-to-Business). Diese Problemstellungen beinhalten auch allgemeine Fragen zum Marshalling/Unmarshalling von Daten und welche Aufwände hierfür notwendig sind, ebenso wie allgemeine Aussagen bzgl. der Echtzeitfähigkeit von CORBA-basierten, verteilten Anwendungen. Die Ergebnisse werden anschließend auf andere Technologien wie COM/DCOM, J2EE und den Web Services, soweit es zulässig ist, übertragen. Die Vergleiche CORBA mit DCOM, CORBA mit J2EE und CORBA mit Web Services zeigen im Detail die Eignung dieser Technologien bzgl. loser und enger Koppelung. Desweiteren werden aus den erzielten Resultaten allgemeine Konzepte bzgl. der Architektur und der Optimierung der Kommunikation abgeleitet. Diese Empfehlungen gelten uneingeschränkt für alle untersuchten Technologien im Zusammenhang mit verteilter Verarbeitung.

A hybrid technology for parallel recording of single ion channels

Hybrid technologies, thanks to the convergence of integrated microelectronic devices and new class of microfluidic structures could open new perspectives to the way how nanoscale events are discovered, monitored and controlled. The key point of this thesis is to evaluate the impact of such an approach into applications of ion-channel High Throughput Screening (HTS)platforms. This approach offers promising opportunities for the development of new classes of sensitive, reliable and cheap sensors. There are numerous advantages of embedding microelectronic readout structures strictly coupled to sensing elements. On the one hand the signal-to-noise-ratio is increased as a result of scaling. On the other, the readout miniaturization allows organization of sensors into arrays, increasing the capability of the platform in terms of number of acquired data, as required in the HTS approach, to improve sensing accuracy and reliabiity. However, accurate interface design is required to establish efficient communication between ionic-based and electronic-based signals. The work made in this thesis will show a first example of a complete parallel readout system with single ion channel resolution, using a compact and scalable hybrid architecture suitable to be interfaced to large array of sensors, ensuring simultaneous signal recording and smart control of the signal-to-noise ratio and bandwidth trade off. More specifically, an array of microfluidic polymer structures, hosting artificial lipid bilayers blocks where single ion channel pores are embededed, is coupled with an array of ultra-low noise current amplifiers for signal amplification and data processing. As demonstrating working example, the platform was used to acquire ultra small currents derived by single non-covalent molecular binding between alpha-hemolysin pores and beta-cyclodextrin molecules in artificial lipid membranes.

Parallel modeling of the electric field distribution in the brain

The term "Brain Imaging" identi�es a set of techniques to analyze the structure and/or functional behavior of the brain in normal and/or pathological situations. These techniques are largely used in the study of brain activity. In addition to clinical usage, analysis of brain activity is gaining popularity in others recent �fields, i.e. Brain Computer Interfaces (BCI) and the study of cognitive processes. In this context, usage of classical solutions (e.g. f MRI, PET-CT) could be unfeasible, due to their low temporal resolution, high cost and limited portability. For these reasons alternative low cost techniques are object of research, typically based on simple recording hardware and on intensive data elaboration process. Typical examples are ElectroEncephaloGraphy (EEG) and Electrical Impedance Tomography (EIT), where electric potential at the patient's scalp is recorded by high impedance electrodes. In EEG potentials are directly generated from neuronal activity, while in EIT by the injection of small currents at the scalp. To retrieve meaningful insights on brain activity from measurements, EIT and EEG relies on detailed knowledge of the underlying electrical properties of the body. This is obtained from numerical models of the electric �field distribution therein. The inhomogeneous and anisotropic electric properties of human tissues make accurate modeling and simulation very challenging, leading to a tradeo�ff between physical accuracy and technical feasibility, which currently severely limits the capabilities of these techniques. Moreover elaboration of data recorded requires usage of regularization techniques computationally intensive, which influences the application with heavy temporal constraints (such as BCI). This work focuses on the parallel implementation of a work-flow for EEG and EIT data processing. The resulting software is accelerated using multi-core GPUs, in order to provide solution in reasonable times and address requirements of real-time BCI systems, without over-simplifying the complexity and accuracy of the head models.

Aktionenlernen mit Selbstorganisierenden Karten und Reinforcement Learning

Die vorliegende Arbeit beschäftigt sich mit der Entwicklung eines Funktionsapproximators und dessen Verwendung in Verfahren zum Lernen von diskreten und kontinuierlichen Aktionen: 1. Ein allgemeiner Funktionsapproximator – Locally Weighted Interpolating Growing Neural Gas (LWIGNG) – wird auf Basis eines Wachsenden Neuralen Gases (GNG) entwickelt. Die topologische Nachbarschaft in der Neuronenstruktur wird verwendet, um zwischen benachbarten Neuronen zu interpolieren und durch lokale Gewichtung die Approximation zu berechnen. Die Leistungsfähigkeit des Ansatzes, insbesondere in Hinsicht auf sich verändernde Zielfunktionen und sich verändernde Eingabeverteilungen, wird in verschiedenen Experimenten unter Beweis gestellt. 2. Zum Lernen diskreter Aktionen wird das LWIGNG-Verfahren mit Q-Learning zur Q-LWIGNG-Methode verbunden. Dafür muss der zugrunde liegende GNG-Algorithmus abgeändert werden, da die Eingabedaten beim Aktionenlernen eine bestimmte Reihenfolge haben. Q-LWIGNG erzielt sehr gute Ergebnisse beim Stabbalance- und beim Mountain-Car-Problem und gute Ergebnisse beim Acrobot-Problem. 3. Zum Lernen kontinuierlicher Aktionen wird ein REINFORCE-Algorithmus mit LWIGNG zur ReinforceGNG-Methode verbunden. Dabei wird eine Actor-Critic-Architektur eingesetzt, um aus zeitverzögerten Belohnungen zu lernen. LWIGNG approximiert sowohl die Zustands-Wertefunktion als auch die Politik, die in Form von situationsabhängigen Parametern einer Normalverteilung repräsentiert wird. ReinforceGNG wird erfolgreich zum Lernen von Bewegungen für einen simulierten 2-rädrigen Roboter eingesetzt, der einen rollenden Ball unter bestimmten Bedingungen abfangen soll.

COBAMOS: Endnutzerprogrammierung auf Basis nachrichtenbasierter Komponentenvernetzung

Die Aufgabenstellung, welche dieser Dissertation zugrunde liegt, lässt sich kurz als die Untersuchung von komponentenbasierten Konzepten zum Einsatz in der Softwareentwicklung durch Endanwender beschreiben. In den letzten 20 bis 30 Jahren hat sich das technische Umfeld, in dem ein Großteil der Arbeitnehmer seine täglichen Aufgaben verrichtet, grundlegend verändert. Der Computer, früher in Form eines Großrechners ausschließlich die Domäne von Spezialisten, ist nun ein selbstverständlicher Bestandteil der täglichen Arbeit. Der Umgang mit Anwendungsprogrammen, die dem Nutzer erlauben in einem gewissen Rahmen neue, eigene Funktionalität zu definieren, ist in vielen Bereichen so selbstverständlich, dass viele dieser Tätigkeiten nicht bewusst als Programmieren wahrgenommen werden. Da diese Nutzer nicht notwendigerweise in der Entwicklung von Software ausgebildet sind, benötigen sie entsprechende Unterstützung bei diesen Tätigkeiten. Dies macht deutlich, welche praktische Relevanz die Untersuchungen in diesem Bereich haben. Zur Erstellung eines Programmiersystems für Endanwender wird zunächst ein flexibler Anwendungsrahmen entwickelt, welcher sich als Basis zur Erstellung solcher Systeme eignet. In Softwareprojekten sind sich ändernde Anforderungen und daraus resultierende Notwendigkeiten ein wichtiger Aspekt. Dies wird im Entwurf des Frameworks durch Konzepte zur Bereitstellung von wieder verwendbarer Funktionalität durch das Framework und Möglichkeiten zur Anpassung und Erweiterung der vorhandenen Funktionalität berücksichtigt. Hier ist zum einen der Einsatz einer serviceorientierten Architektur innerhalb der Anwendung und zum anderen eine komponentenorientierte Variante des Kommando-Musters zu nennen. Zum anderen wird ein Konzept zur Kapselung von Endnutzerprogrammiermodellen in Komponenten erarbeitet. Dieser Ansatz ermöglicht es, unterschiedliche Modelle als Grundlage der entworfenen Entwicklungsumgebung zu verwenden. Im weiteren Verlauf der Arbeit wird ein Programmiermodell entworfen und unter Verwendung des zuvor genannten Frameworks implementiert. Damit dieses zur Nutzung durch Endanwender geeignet ist, ist eine Anhebung der zur Beschreibung eines Softwaresystems verwendeten Abstraktionsebene notwendig. Dies wird durch die Verwendung von Komponenten und einem nachrichtenbasierten Kompositionsmechanismus erreicht. Die vorgenommene Realisierung ist dabei noch nicht auf konkrete Anwendungsfamilien bezogen, diese Anpassungen erfolgen in einem weiteren Schritt für zwei unterschiedliche Anwendungsbereiche.