Thesis (PH. D.)--Johns Hopkins university, 1915. / Biography.

Thesis (Ph. D.)--Florida State University, 2006. / Advisor: Frances Prevatt, Florida State University, College of Education, Dept. of Educational Psychology and Learning Systems. Title and description from dissertation home page (viewed June 8, 2006). Document formatted into pages; contains x, 158 pages. Includes bibliographical references.

Theses (Ed.S.)--Marshall University, 2005. / Title from document title page. Includes abstract. Document formatted into pages: contains 25 p. Bibliography: p. 24-25.

University, Diss., 2003--Dortmund.

Universiẗat, Diss., 2005--Stuttgart.

Doct. diss.--Department of psychology--Stockholm university, 1999. / Ouvrage constitué de cinq études indépendantes, contenant des résumés en anglais, en français, en espagnol. Bibliogr. en fin de chaque étude.

Cette thèse vise l’étude de nouvelles architectures d’auto test pour les convertisseurs de type pipeline. En production, les convertisseurs sont testés en fonctionnement statique et dynamique. Les techniques de test statique de linéarité sont les techniques les plus coûteuses durant la phase de production. La mesure des performances statiques utilise un stimulus à haute linéarité et très basse fréquence et la méthode de l’histogramme, nécessitant la collecte d’un grand nombre d’échantillons en sortie afin de moyenner le bruit. Ainsi, la quantité de données nécessaire augmente exponentiellement avec la résolution du CAN sous test. Pour cette raison, la réduction du temps de test des CANs est un domaine de recherche qui attire de plus en plus d’attention. Récemment, des nouvelles solutions ont été mises au point pour réduire de façon importante le temps de test, mais aucune solution d’auto test considérant un générateur de signal de haute résolution en combinaison avec une technique d'analyse intégrée, réduisant considérablement la quantité de données, n’a encore été développée. Dans le cadre de cette thèse, on envisage l’étude de techniques d’auto test statique pour ce type de convertisseurs. En particulier, cette thèse présente un générateur de stimulus de test intégré à haute linéarité et une technique modifiée de servo-loop qui, en combinaison avec un algorithme de test de linéarité avec réduction de codes, conduit à la définition d'une stratégie efficace et précise de test intégré pour les CANs de type pipeline. La thèse inclut la validation expérimentale des techniques proposées, en coopération avec ST Microelectronics, Grenoble. / This PhD thesis is aimed at exploring new Built-In-Self-Test (BIST) techniques for static linearity characterization of pipeline ADCs. During the production phase, the static and dynamic performances of the ADCs are tested. Static linearity test techniques are one of the more expensive test procedures that are performed at production line. The measurement of the static linearity performance requires the application of a low frequency high linearity stimulus and the collection of a high volume of output samples for noise averaging, usually using a histogram-based test setup. Thus, as the resolution of state-of-the-art ADCs increases, test time for static linearity characterization increases exponentially. For this reason, the reduction of the ADC test time is a hot topic that has gained an increasing interest over the past years. New techniques have recently been proposed to effectively reduce test time, but no BIST technique has yet been developed that considers a high resolution signal generator in combination with an on-chip analysis technique that dramatically reduces the amount of data. In this thesis, static linearity BIST techniques will be investigated for pipeline ADCs. In particular, this thesis presents a novel high-linearity on-chip test stimulus generator and a modified servo-loop technique that, in combination with reduced-code linearity test algorithms, lead to the definition of an efficient and accurate BIST strategy for pipeline ADCs. The work includes the experimental validation of the proposed techniques in collaboration with STMicroelectronics, Grenoble.

Thesis (M.S.)--Boston University

Testing is a part of the software development life cycle (SDLC) which ensures the quality and efficiency of the software. It gives confident to the developers about the system by early detecting faults of the system. Therefore, it is considered as one of the most important part of the SDLC. Unfortunately, testing is often neglected by the developers mainly because of the time and cost of the testing process. Testing involves lots of manpower, specially for a large system, such as distributed system. On the other hand, it is more common to have bugs in a large system than a small centralized system and therefore there is no alternative of testing to find and fix the bugs. The situation gets worst if the developer follows one of the most powerful development process called continuous integration process. This is because developers need to write the test cases in each cycle of the continuous integration process which increase the development time drastically. As a result, testing often neglected for large systems. This is an alarming situation because distributed system is one of the most popular and widely accepted system in both industries and academia. Therefore, this is one of the highly pressured areas where lot of developers engaged to provide distributed software solutions. If these systems delivered to the users untested, there is a high possibility that we will end up with a lot of buggy systems every year. There are also a very few number of testing framework exist in the market for testing distributed system compared to the number of testing framework exists for traditional system. The main reason behind this is, testing a distributed system is far difficult and complex process compares to test a centralized system. Most common technique to test a centralized system is to test the middleware which might not be the case for distributed system. Unlike the traditional system, distributed system can be resides in multiple location of different corners of the world. Therefore, testing and verification of distributed systems are difficult. In addition to this, distributed systems have some basic properties such as fault tolerance, availability, concurrency, responsiveness, security, etc. which makes the testing process more complex and difficult. This research proposed a multi-agent based testing framework for distributed system where multiple agent communicate with each other and accomplish the whole testing process for a distributed system. The bullet proof idea of testing centralizes system has been reused partially to design the framework so that developers will be more comfortable to use the framework. The research also focused on the automation of testing process which will reduce the time and cost of the whole testing process and relief the developer from re-generating the same test cases over and over before each release of the application. This paper briefly described the architecture of the framework and communication process between multiple agents.

No description available.