This paper deals with speech recognition, its usability, expandability, techniques and models which are used in speech recognition applications. The main criteria of the expandability is the possibility to develop own additional features. At first the paper focuses on the different state of the art techniques such as Dynamic Time Warp, Hidden-Markov-Method, neural networks or the statistical speech recognition model. In addition actual speech recognition applications will be tested. The test focus is on the usability (How easy can the software be installed and used? How good is the speech recognition?) and expandability (Does the software provide programming interfaces to extend the software with custom features?) Speech recognition should not only help us using our voice instead of the keyboard to write words faster, or help computer gamers to perform complex key combinations with single words commands, it should also help handicapped people to interact with their surrounding environment. It provides the possibility to control a PC with spoken words and phonemes. Due to the possibility to train the speech recognition system software to the voice of the user, the program adapts to the user instead of the user to the program.

This bachelor thesis deals with the practical development of a prototype software plugin for the speech recognition application Simon Listens. The extension of Simon Listens shall provide the possibility to access the KNX bus, a worldwide standard in home automation. Furthermore it provides the possibility to access and configure sensors or actuators distributed in a house, building or office blocks. Possible actuators are for example: light control units, blinds, heater or airconditioning units. The plugin’s purpose is to help elderly people to control their homes by voice. The bachelor thesis “speech recognition in home automation” from this author is used as basis for this paper, and describes the speech recognition models and evaluates speech recognition applications as a basis for implementing the prototype. In the first few chapters the paper introduces the theoretical background of the KNX systems and describes the basics needed to access a KNX bus system. It covers KNX itself, KXN/IP (the transformation from KNX bus system to IP), EIBD (a computer interface to send KNX/IP telegrams), and the ETS (the software for the configuration of the KNX actuators and sensors). The second part deals with the practical development which explains the implementation of the Simon Listens plugin. It starts with the basic requirements which are needed to setup the development environment and the required application programs which provide access to the KNX bus. Afterwards the successful implementation of the prototype is explained.

97% of people over 60 and 88% of people aged 80 and more live in a private household. Most of them want to stay there as long as possible. Offering these people a dignified life and autonomy in daily-living is a big challenge. Caregivers are often not affordable. As a result of an increasing ratio of elderly people, known as demographic change, it will be even harder in future. Ambient Assisted Living (AAL) provides technical solutions for this problem. Existing technologies are used to support people with impairments and the elderly wherever it is possible. Indoor localization systems are a promising solution to assist people at home. Knowing the position of a resident enables systems like automatic light control and navigation of emergency services. The aim is to increase peoples’ quality of life, autonomy and safety. There are a lot of different localization systems on the market. Depending on the used technique they differ in range, accuracy, costs and type of signal, e.g. infrared, ultrasonic, radio signals or visible light. Some studies compared these systems and tried to evaluate them. Very often, accuracy became the most important characteristic in these surveys. This thesis offers a survey of indoor localization systems in view of suitability for AAL applications, which require other priorities, such as lower costs being more important than high accuracy. To find these requirements, studies about the desires and needs of elderly people are analyzed. Information about localization schemes, e.g. Time of arrival (TOA) and Received signal strength (RSS) are presented, which help to find the appropriate way of localization. On the basis of the requirements, technologies like Bluetooth, WiFi, RFID and UWB are reviewed afterwards.

Most of the current KNX clients, whether PC or cell phone based, are proprietary. This thesis describes the development of an open source Android App for connecting to KNX networks using an IP connection (wireless or LAN cable). The Android App uses the open source Java Camilero NG library that implements the KNXnet/IP tunneling protocol using UDP as the transport. Background material, preliminary tests, and future development are covered.

The University of Applied Sciences Technikum Wien is expanding into the areas of Ambient Assistive Technologies and Smart Homes. KNX is one of the key technologies in this area. For current and future KNX projects, students will need a good basic understanding of KNX in order develop new fieldbus devices and software applications. This paper is a how-to guide about using the current KNX resources provided by the University.

The purpose of this paper is to document the development of a client program for the Beckhoff TwinCAT system. The installation and the configuration of both the TwinCAT and the ADS API library are covered. A detailed guide for compiling and installing a PLC program is provided. Both the PLC variables and hardware pins are accessed by a .NET client program to demonstrate the various API calls for reading, writing, and receiving notifications. Finally, the remote communication interfaces between the TwinCATs as well as using a web service are demonstrated.

This paper gives an overview of the current available media center software packages, which can be used for integrating functions of home automation. The principle idea is on the one hand to have the possibility of using the media data which are available in the meanwhile in each household and on the other hand to switch on and off the light, for example by pressing just a button on an infrared remote control. In the best case the interface of the media center for controlling the home automation is exactly the same as it is used for all other functions like viewing pictures and movies or just listening to stored music files. Within this scenario the available TV set should be the central gateway to all functions provided by the media center. In the first step the requirements for such a media center will be defined and in the following chapters the listed media centers will be evaluated. The overall aim is to find out, which software package can be used easily to fulfill the defined requirements and provides the possibility to integrate a dedicated function handling the connection to the home automation system. As the physical medium for the home automation system the KNX bus was chosen because of its widespread use and the interfacing possibilities which can be done with two different packages called eibd or calimero. Those two programs provide an interface to the KNX bus on different ways. The eibd is a daemon which can be run under any Linux system and provides a shared memory area which can be accessed to send commands to the connected KNX bus system. On the other hand the calimero software is a tool programmed under Java (which makes it possible to be used under Windows and under Linux) which handles the KNX accesses. So by finishing this paper one of the described media center solutions in cooperation with eibd or calimero has to be defined and used in a second part for the real integration of the needed functions for accessing the home automation system via a media center. At the end of this paper we will see, that the mentioned media center called XBMC in cooperation with the daemon eibd for accessing the KNX bus seems to be the most sufficient solution for the defined application which has to be developed in a second step.

This paper describes the implementation of a home automation system into the user interface and program structure of a media center software. The principle idea is that the user is capable of controlling his home from one central point. Since a television set is available in nearly every home it would be the best approach to use this equipment for displaying the content and giving the user the possibility of initiating different actions on the home automation bus. Within this document the details of the implementation of the extension modules and the integration afterwards are described. Each developed module is explained in detail which means that finally this thesis can be used as a basic documentation for further development and better understanding of the functionality. As part of a first paper the author decided to use the XBMC media center software for extension with the needed functionality and the usage of the KNX-bus for home automation systems. Further details and descriptions how decisions have been made can be seen under [1]. The used system for the prototype development consists of the media center software called XBMC, extended by the self developed packages for controlling and visualizing the KNX bus and the so called eibd which is a software daemon for Linux connecting any computer with the KNX network. Communication between the eibd and the extension modules is based on sockets over the available network infrastructure. When the eibd receives a command via the socket connection it translates it to commands for being sent out on the connected KNX-bus. At the end it can be stated that the proof of concept was successful and the implementation and integration of a home automation system, based on a KNX-bus, into the XBMC media center can be done in the described manner. The current available prototype software has to be extended for seamless use in the field but the principle functionality was implemented and is available for being used to control a smart home.

Die vorliegende Master Thesis findet im Rahmen des "Mobile Motion Advisor"-Projektes (MMA) statt. Das Ziel des genannten Projektes ist, dass SportlerInnen bei der Uebungsdurchfuehrung durch ExpertInnen ein Feedback bekommen. So werden beispielsweise biomechanische oder physiologische Parameter an Ort und Stelle mit einem mobilen Endgeraet (Smartphone) erfasst und an einen Server zur Analyse uebertragen. Die Klassifikation der Daten und Generierung des (unmittelbaren) Feedbacks basiert auf dem Wissen von ExpertInnen unterschiedlicher Disziplinen der Sportwissenschaften und wird sportartspezifisch adaptiert. Durch die Aufzeichnung der Daten am Server koennen diverse Trainingseinheiten miteinander verglichen werden und fuer eine Analyse und Bewertung kann von jedem Rechner mit Internetzugang zugegriffen werden. In der vorliegenden Arbeit wird die Software fuer eine bestehende Sensorplattform entwickelt. Diese Plattform dient dem Erfassen der diversen sportspezifischen Sensorwerte, die gegebenenfalls aufbereitet werden, und der Uebertragung dieser Daten ueber eine drahtlose Schnittstelle (ANT) an ein Smartphone. Die Weiterleitung der Daten an den Server sind aber nicht Teil dieser Arbeit. Als erste Sportbereiche werden Mountainbiken und Kraftkammer behandelt. Fuer die Verwendung als Sensorplattform gibt es Echtzeitforderungen, aber auch einige Komponenten, die ein asynchrones, undeterministisches Verhalten aufweisen, weshalb die Echtzeitforderungen mithilfe einer geeigneten Systemarchitektur erfuellt werden muessen. Desweiteren erfordern einige Sensoren eine entsprechende Signalverarbeitung. Auch der maximale Datendurchsatz der Funkverbindung ist zu gering um in manchen Applikationen alle Daten zu uebertragen. Kern der Systemarchitektur stellt zum Einen ein Scheduler dar, welcher auf statischem Scheduling basiert und zum Anderen die Auslagerung der asynchronen Ereignisse. Fuer die Filterung musste eine Loesung gefunden werden, die keine Phasenverschiebung erzeugt, weshalb ein akausales Prinzip benutzt wird. Da die Datenrate der Funkverbindung nicht geaendert werden kann, wurde versucht Messdaten moeglichst zusammenzufassen und bei Daten, die hoeherfrequent abgetastet werden mussten, wurden nur die Gefilterten uebertragen. Zusaetzlich werden aber auch alle Roh-Daten auf eine SDKarte geschrieben. Am Ende erfolgt die Evaluierung der implementierten Loesungen. Abschliessend erfolgt noch ein Ausblick und einige Verbesserungsvorschlaege werden diskutiert.