Unless stated otherwise, all topics are offered at the Alpen-Adria-Universität Klagenfurt at the group of Prof. Wilfried Elmenreich.
In this thesis, FREVO should be extended by a functionality that tests a given algorithm by evolving adversary behavior in the evironmental simulation that is used to test the quality of the solution. This can range from a simple tryout of rendom seeds for the environment to find the one where the performance is worse up to applying a search algorithm (probably reusing the FREVO component for optimizing the candidate here) to find the configuration that is most problematic for the previously found algorithm.
Introducing a Criteria for Measuring the Complexity of Environments in terms of Robotic Exploration
Type: Master thesis
Type: Master thesis
Advisor: Khalil Al-rahman Youssefi Darmian
The complexity of the environment has a significant impact on the performance and quality of algorithms designed for mobile robots, such as Search methods and Coverage Path Planning algorithms. To the extent that algorithms can even be categorized based on their performance in different environments. In this project, the aim is to provide a suitable and reliable method for measuring the complexity of environments. This degree of complexity can be "relative" or "absolute".
For example, in the figure below, environments 1 to 3 have the same shape (i.e., equal Area and Perimeter). Intuitively, Environment 3 is more complex than environment 2, and environment 2 is more complicated than environment 1. However, in science, it is preferred to use numbers rather than intuitive reasoning. Thus, the question is, how different is the difficulty of these environments?
For example, in the figure below, environments 1 to 3 have the same shape (i.e., equal Area and Perimeter). Intuitively, Environment 3 is more complex than environment 2, and environment 2 is more complicated than environment 1. However, in science, it is preferred to use numbers rather than intuitive reasoning. Thus, the question is, how different is the difficulty of these environments?
Framework for evolutionary design and optimization tasks
Type: Master thesis
Advisor: Wilfried Elmenreich
FREVO is an open-source framework developed in Java to help engineers
and scientists in evolutionary design or optimization tasks. The major
feature of FREVO is the componentwise decomposition and separation of
the key building blocks for each optimization tasks. We identify these
as the problem definition, solution representation and the optimization
method. This structure enables the components to be designed separately
allowing the user to easily swap and evaluate different configurations
and methods or to connect an external simulation tool.
The intended work for the thesis can be done in one of the following two
ways:
(i) implementation of a new problem domain within the framework and
evaluate how well the neural controllers can be evolved to solve the
task.
(ii) implementation a new representation for the controller (e.g.
Spiking Neural Networks) and apply it to existing problem domain
implementations.
Co-Evolve Adversaries in FREVO
Type: Master thesis
Advisor: Wilfried Elmenreich
FREVO
is an open-source framework developed in Java to help engineers and
scientists in evolutionary design or optimization tasks. The major
feature of FREVO is the componentwise decomposition and separation of
the key building blocks for each optimization tasks. FREVO is used with the EU H2020 project CPSwarm.In this thesis, FREVO should be extended by a functionality that tests a given algorithm by evolving adversary behavior in the evironmental simulation that is used to test the quality of the solution. This can range from a simple tryout of rendom seeds for the environment to find the one where the performance is worse up to applying a search algorithm (probably reusing the FREVO component for optimizing the candidate here) to find the configuration that is most problematic for the previously found algorithm.
Implementation of a Communication Network for Swarm Robotics
Type: Bachelor thesis
Advisor: Wilfried Elmenreich
Communication plays a crucial role for an implementation of successful behaviour in the field of swarm robotics. The goal of the project is an implementation of a communication network for a large colony of robots. In the project you will set up and program an ESP8266 WiFi module for an Arduino system.
Communication plays a crucial role for an implementation of successful behaviour in the field of swarm robotics. The goal of the project is an implementation of a communication network for a large colony of robots. In the project you will set up and program an ESP8266 WiFi module for an Arduino system.
Design of a Bootloader for Programming of Robots in Swarm Applications
Type: Bachelor thesis or Research project
Advisor: Sergii Zhevzhyk, Wilfried Elmenreich
Advisor: Sergii Zhevzhyk, Wilfried Elmenreich
Robotic swarms consist of many small robots accomplishing some complex tasks. Real experiments with such large colonies may be time consuming when some modifications of programs are needed. The idea of the project is to develop a bootloader for autonomous robots in order to simplify uploading of new or modified programs. The bootloader should communicate with autonomous robots via an ESP 8266 WiFi module and offer simple and understandable interface for uploading of a sketch.
Fitness landscape analysis
Type: Bachelor thesis
Advisor: Wilfried Elmenreich
For a set of problems, the fitness landscape should be analyzed using a simulation approach. You should have a fair knowledge of Java and be interested in simulation.
For a set of problems, the fitness landscape should be analyzed using a simulation approach. You should have a fair knowledge of Java and be interested in simulation.
Setting up a robot simulator
Type: Research project
Advisor: Wilfried Elmenreich
Look into existing simulators and compare them based on their features.
No comments:
Post a Comment