· · · Institutional Repository

This thesis investigates how methods for collaborative agents can be applied to the AIBO (Artificial Intelligence roBOt) of SONY Corporation. A method for collaboration will be explained in detail, and the design and implementation of a working prototype will be presented. The software works in an ad-hoc wifi network with clients that can connect and disconnect at all times. Data can be shared between the clients (AIBOs) so that they can determine the right objectives for themselves and fine-tune their actions with the others.

The methodologies ARIS and DEMO combined on a theoretical level, at the Corporate Dienst of Rijkswaterstaat.

Ter afsluiting van de bachelor TI is in het bachelorproject een iPhone-webapplicatie ontworpen voor Greetinq Voicemail.

There exist several formalisms for representation and reasoning in dynamic systems, for example, Dynamic Influence Diagrams (DID), Influence Diagrams (ID), Dynamic Bayesian Networks (DBN), Bayesian Networks (BN), Hidden Markov Models (HMM), Markov Decision Processes (MDP), and Partially Observable Markov Decision Processes (POMDP). All these formalisms belong to graphical models based on probability theory. It has been shown that all probability models can be seen as variants of one generalization model. The purpose of this thesis is to review these models, to try to propose a unifying representation of these models at some generalization level (assuming DID level), and to test them in practice.

At the Man Machine Interaction research group at the Delft University of Technology research is being done on the subject of aggression detection in trains. The goal of this project is to research different aspects of train surveillance, including video surveillance, but also audio surveillance storyboard based modeling. This thesis discusses the current state of the art methods and techniques that or being applied, or could be applied to the task of automated video surveillance. This work discusses the application to the video surveillance problem of several methods, most notably motion detection, face tracking, face recognition and facial expression analysis.

The project addresses cost savings and productivity improvement by pooling of resources between independent companies through multi-agent, bargaining-based decision methods. Cost savings are realized by optimizing plans across companies, while the pooling decision and the benefit distribution is done by bargaining. Applications with these characteristics are wide-spread including logistics (pooling of fleets) and manufacturing (joint manufacturing plans). This thesis comprises the project proposal as a research project submitted to the Sixth Framework Programme, Information Society Technologies, and the realization of the proof-of-concept. This proof is provided by the design and implementation a prototype optimizer for the Capacitated Vehicle Routing Problem (CapVRP) in order to simulated pooling versus non-pooling strategies based on a – simplified – business case. The optimizer uses the genetic algorithm meta-heuristic. Extensive experimentation using published CapVRP instances and benchmarks leads to the delivery of the proof-of-concept.

New ideas to improve automatic speech recognition have been proposed that make use of context user information such as gender, age and dialect. To incorporate this information into a speech recognition system a new framework is being developed at the MMI department of the EWI faculty at the Delft University of Technology. This toolkit is called Gaia and makes use of Dynamic Bayesian networks. In this thesis a basic speech recognition system was built using Gaia to test if speech recognition is possible using Gaia and DBNs. DBN models were designed for the acoustic model, language model and training part of the speech recognizer. Experiments using a small data set proved that speech recognition is possible using Gaia. Other results showed that training using Gaia is not working yet. This issue needs to be addressed in the future and also the speed of the toolkit.

In this thesis a research about a face recognition system is presented. The system uses a database of images to identify individuals. The starting point is a huge database of “known faces”, the FERET database. If a new face is presented the system has to decide whether this face is a member of the database or not. The matching of a face can be realized using the eigenfaces algorithm. Applying PCA (Principal Component Analysis) the database can be reduced to a finite number of eigenfaces, such that every face can be approximated by a weighted sum of eigenfaces. Every face can be represented by a column of elements and a distance measure is used to compute the distance between columns. In this thesis a face recognition system has been implemented and tests are performed. Two algorithms PCA (Principal Component Analysis) and ICA (Independent Component Analysis) have been implemented and the test results are compared with each other.

Many public or open areas are facilitated with cameras at every angle to monitor the security of that area for keeping the citizens safe, which is known as the surveillance system. For this moment, the best solution to approach a safety environment has to be done by human. Even though a human is the most intelligent creature in this world, still there are some shortcomings from the existing way. Because of these kinds of shortcomings, human keeps finding new discoveries to replace them and make the best of it. In order to support this surveillance system, a recognition and tracking system is built to detect an abandoned luggage in the public transportation area such as train central station and airport. The goal of this project is to design and implement an algorithm which will be able to detect abandoned luggage using the captured images or videos from the camera as the input of the system. The algorithm realizes image segmentation and image tracking, creates blobs of objects, labels the blobs and finally gives warning when an abandoned luggage is detected. Also a database is developed to store all the media data.

Major research efforts in Bioinformatics include sequence alignment, gene finding, genome assembly, protein structure alignment, protein structure prediction, prediction of gene expression and protein-protein interactions, and the modeling of evolution. To perform these specific tasks different tools are used. Using these tools separately is a time consuming, inefficient and expensive process. MIDAS is a tool that integrates sequence alignment, genome annotation, and spectral clustering and alignment under the same application. The challenge in this project is in representing the knowledge and analyzing the genome data. The DNA data is first transformed into Fourier domain and clustered in MATLAB based on Euclidean distances between the sequences. Our tool allows visualizing the DNA spectra together with a hierarchical tree in a multimodal interface. This in turn enables a bioinformatician to analyze patterns of a group of sequences. MIDAS is a standalone application which provides an interface around standard sequence alignment tools such as BLAT, ClustalW, as well as newer alignment tools such as Spectrogram analysis via integrating MATLAB code, server connections and data visualizations. JAVA is used as the main programming language during the development of MIDAS. The Spectrogram Analysis script-files in MATLAB are converted into JAVA classes. These classes are used to run standalone MATLAB applications from within JAVA.

In slaap vallen in het verkeer is nog steeds een grote oorzaak van verkeersongevallen en leidt elk jaar nog tot veel slachtoffers. Een systeem dat kan helpen om deze ongevallen te voorkomen zou derhalve erg nuttig zijn. Het doel van het project B.O.O.S. is om een systeem te ontwikkelen dat kan detecteren wanneer de bestuurder van een auto in slaap aan het vallen is en hem daarop te attenderen door middel van een geluidssignaal. Om te detecteren of iemand in slaap aan het vallen is wordt door het systeem continu gekeken naar de status van de ogen en wordt er aan de hand van de resultaten beredeneerd of er actie ondernomen moet worden. Dit project laat zien dat het goed mogelijk is om een systeem te maken dat kan detecteren wanneer iemand in slaap valt en actie kan ondernemen om zo (dodelijke) ongelukken in het verkeer te voorkomen.

In the past few years, traffic incident management has been a subject of interest in both politics and research. After an incident has occurred, incident management measures are intended to clear the motorway for traffic, inform other road users and treat victims as soon as possible. These measures largely constitute formal agreements, procedures and coordination efforts between all parties involved. Although the new incident management measures currently in effect in the Netherlands have improved the time required to clear an incident scene, information technology is expected to decrease the required time even further by improving communication and situational awareness and supporting analysis and evaluation of the incident response. In this scope, a new IT system is developed, called the Incident Management Information and Communication System (IMICS). Based on an existing framework, the Java Agent DEvelopment framework (JADE), IMICS offers a blackboard like functionality through which the police, fire brigade, ambulance, traffic control centres, the shared control rooms and the Department of Public Works can update and share incident and task information both at the control rooms and at the incident scene. In order to provide clear and unambiguous information, an ontology and an information management system have been designed that closely follow the incident management procedures. Finally, a part of the IMICS design has been implemented and tested in a lab setting as a proof of concept. The results show IMICS improves availability and reliability of data in a lab setting.

Human machine interaction is not as natural as interaction among humans. That is the reason why is not yet possible to completely substitute face-to-face communication by human-machine interaction in spite the theoretical feasibility of such a substitution in several professional areas including education and certain medical branches. To approach the naturalness of face-to-face interaction, machines should be able to emulate the way humans communicate with each other. Multi-media and man-machine communication systems could promote more efficient performance if machine understanding of facial expressions is improved. Based on this problematic issue, the current thesis is focused on the recognition of emotional facial expressions, finding interrelation between the facial expressions and labels and finally the classification of the expressions. One of the parts of this project is an emotional database which will contain images of faces, their corresponding Action Units and their labels. The contribution of this database to the problem stated above is that it can be used by systems in order to recognize emotional facial expressions given one of the database data i.e. action units’ combination. The other part of the project, which is an expert system for emotional classification, will enable to classify emotional expressions, the ones included in the database and all the possible combination of them.

Major research efforts in Bioinformatics include sequence alignment, gene finding, genome assembly, protein structure alignment, protein structure prediction, prediction of gene expression and protein-protein interactions, and the modeling of evolution. To perform these specific tasks different tools are used. Using these tools separately is a time consuming, inefficient and expensive process. MIDAS is a tool that integrates sequence alignment, genome annotation, and spectral clustering and alignment under the same application. The challenge in this project is in representing the knowledge and analyzing the genome data. The DNA data is first transformed into Fourier domain and clustered in MATLAB based on Euclidean distances between the sequences. Our tool allows visualizing the DNA spectra together with a hierarchical tree in a multimodal interface. This in turn enables a bioinformatician to analyze patterns of a group of sequences. MIDAS is a standalone application which provides an interface around standard sequence alignment tools such as BLAT, ClustalW, as well as newer alignment tools such as Spectrogram analysis via integrating MATLAB code, server connections and data visualizations. JAVA is used as the main programming language during the development of MIDAS. The Spectrogram Analysis script-files in MATLAB are converted into JAVA classes. These classes are used to run standalone MATLAB applications from within JAVA.

Dutch freeways suffer from severe congestion during rush hours or incidents. Research shows that 64% of congested traffic during rush hour consists of commuter traffic [30]. A traffic congestion increases travel time, resulting in a delay for travelers. Reliable travel time predictions are essential for Dynamic Routing, in which travelers can be rerouted to avoid congestions. Travel times can be calculated from vehicle speed [41] in case of free flowing traffic. In case of congestion, we will make an estimation error regarding the travel time. Therefore, an accurate speed prediction model is necessary. In this thesis, the predictability of the average vehicle speed by Bayesian Networks is investigated. A case study is conducted where several Bayesian Network models we propose are evaluated for a well known traffic bottleneck in the Netherlands. We show that Bayesian Networks are capable of predicting the start or end of a congestion at the bottleneck reasonably accurate for a prediction horizon until 30 minutes. Further, we propose a prediction model based on historical data, which is able to predict the average vehicle speed at the bottleneck location for longer prediction horizons. In the end, we propose a hybrid model which combines our Bayesian Network and our prediction model based on historical data. This hybrid model is able to predict a traffic congestion with an accuracy of 85% for a prediction horizon of 2.5 hours. The results of our case study show that modeling traffic using Bayesian Networks is promising. Our models can form the input for a travel time prediction model for Dynamic Routing.