In this thesis, we present an approach to integration of case-based reasoning and Bayesian reasoning for decision support. Our design is meant to provide physicians with decision support in the context of palliative care for lung cancer patients. Because of delays in the medical data, we created an intermediate application with the aim to assist people in choosing an adequate wine for a given meal. We have developed a system that is able to utilize both the general knowledge of the Bayesian network and the specialized knowledge of the case base. Our results shows that the combination of CBR and BN are able to discover solutions that would not been found by using only one of the methodologies.

In this report I propose a method of navigation for differentially wheeled robots inspired by path integration in certain social insects like bees and ants. It is a very simple method, intended for use in low-tech robots with very limited hardware, such as swarm robots. Path integration is essentially dead reckoning as used by animals, calculating the relative position based on the movements made since the last known position. It is a tried and true method of navigation that also has significant flaws, especially in that inaccuracies accumulate and magnify over time. In this report I want to examine whether communication and information sharing between robots in a swarm can alleviate some of the drawbacks, and make it a viable method for navigation for swarm robots over relatively short distances.

In the later years robotics has seen a huge increase within domestic use, and have now become an affordable tool in the daily life of most people. The goal of this project was to investigate the differences between a physical and virtual robot in terms of increased content knowledge, learning motivation, and interest in science, technology, engineering and mathematics (STEM). To investigate this we conducted an experiment at Trondheim’m International School (THIS), using a quasi-experimental setup with two treatment group, virtual and physical robot. The results showed that there does not exist a statistically significant difference in content knowledge gain, motivation or interest between the robotics group and the simulator group.

In the later years robotics has seen a huge increase within domestic use, and have now become an affordable tool in the daily life of most people. The goal of this project was to investigate the differences between a physical and virtual robot in terms of increased content knowledge, learning motivation, and interest in science, technology, engineering and mathematics (STEM). To investigate this we conducted an experiment at Trondheim’m International School (THIS), using a quasi-experimental setup with two treatment group, virtual and physical robot. The results showed that there does not exist a statistically significant difference in content knowledge gain, motivation or interest between the robotics group and the simulator group.

The goal of the project in this thesis is to explore the viability of an approach to user adaptation where the application context is significantly more constraining than in most cases seen in previous academic work.The project describes and implements a system for rolling out product features incrementally in an optimal way, based on feature adoption statistics within user segments. In other words, the described system should allow for simple personalization of the product.When analyzing the adoption rate of new features, we find that there are indeed clear differences between the identified user segments. However, limitations due to the lack of stable user identity make adaptation approaches based on these data unfeasable in practice.

Thèses (M.Sc.A.)--Université de Sherbrooke (Canada), 2000. / Titre de l'écran-titre (visionné le 20 juin 2006). Publié aussi en version papier.

Stuttgart, Univ., Diss., 2004.

Universiẗat, Diss., 2004--Stuttgart.

Fertigungssysteme für die individualisierte Produktion erfordern Arbeitsabläufe, die von einzelnen Objekten abhängig sind. Maschinelles Lernen (ML) bietet die Möglichkeit, verschiedene Objekte mit Hilfe eines neuronalen Netzes zu klassifizieren. Abhängig von den Klassifikationsergebnissen können Entscheidungen für den nachfolgenden Produktionsschritt getroffen werden. Es wird untersucht, ob es möglich ist, ein neuronales Netz zur Bilderkennung in Echtzeit und in Koordination mit den Maschinen- und Bewegungssteuerungsaufgaben auszuführen. In dieser Arbeit wird die Umsetzung und Messung mit Hilfe einer SPS-Laufzeitumgebung auf einem Standard-Industrie-PC durchgeführt. Die Ausführungszeiten verschiedener Methoden zur Implementierung neuronaler Netze werden gemessen und verglichen. Das schnellste neuronale Netz benötigt eine durchschnittliche Ausführungszeit von nur 39 μs. Darüber hinaus werden die Eigenschaften der verschiedenen Methoden in Bezug auf das Training und die Implementierung der neuronalen Netze innerhalb verschiedener industrieller Steuerungen diskutiert.

La rupture des pentes et l’éclatement des roches, qui représentent deux types de risques naturels fréquents dans le monde, peuvent engendrer des conséquences économiques importantes et des pertes en vie humaine. Malgré que les phénomènes soient étudiés depuis de longues années, il reste encore des questions ouvertes et sans réponse et il est donc encore nécessaire poursuivre les recherches sur cette thématique. Le présent travail de thèse est consacré la modélisation numérique de la stabilité des grandes pentes et de l’éclatement des massifs rocheux en utilisant des méthodes basées sur l’intelligence artificielle en proposant des modifications et des améliorations de telles méthodes. En se basant sur des observations de déplacements de terrain, le glissement de terrain, qui est phénomène commun de la rupture de pentes, est étudié par le processus de Gauss afin de prédire son apparition temporelle. Ensuite, la question d’évaluation de la stabilité des pentes est abordée en utilisant la stratégie de machine à vecteurs de pertinence (RVM) avec des hyper-paramètres adaptatifs. Une approche itérative est proposée afin de déterminer les valeurs optimales des hyper-paramètres. Afin d’améliorer la prédiction, l’évaluation complète de la stabilité des pentes est réalisée en proposant un modèle basé sur la théorie de flou (CM) associé à un processus analytique d’hiérarchisation pondérée (WAHP). Ce modèle est utilisé à l’évaluation de la stabilité de la pente de rive gauche de la centrale hydroélectrique de Jinping 1, dans la région Sud-Ouest de Chine. Enfin, dans la dernière partie, la problématique de l’éclatement des massifs rocheux est abordée en utilisant des modèles basés sur la théorie du flou, en se basant sur une synthèse de 164 cas réels. Des comparaisons entre les résultats numériques et des données de terrain sont présentées pour de différents cas étudiés dans cette thèse. / Slope failures and rock burst, which are two typical types of geological hazards, create tremendous economic damages and cause massive losses to the health of domestic humans or animals every year throughout the world. The severe situation implies that they are still in need to be further studied despite the fact that they have been discussed for a long time. The present work is devoted to presenting the analysis of slope failures and rock burst using some computational intelligent models with modifications and improvements. Then landslide, a common type of slope failure, is analyzed for time occurrence prediction using the Gaussian Process by means of field-observed displacement series. After that, the problem of slope stability evaluation is discussed using the strategy of relevance vector machine (RVM) with adaptive hyper-parameter. An iteration approach is presented to find optimal hyper-parameter values in this chapter. Afterwards, the comprehensive evaluation of slope stability is carried out with the cloud model (CM) and weighted analytical hierarchy process (WAHP) closely related to the left abutment slope of Jinping 1 Hydropower Station, southwest of China. Finally, prediction of rock burst classification is engaged using the cloud models synthesized with the attribution weights on the basis of 164 rock burst cases. In each modeling of the associated problems, comparisons are given on the performance of each strategy as well as some evaluations.