Practicals & Theses
List of practicals and theses at the institute (see also TI Research Presentations)
Details for Dynamic Reconfiguration Framework
The aim of this project is to develop a dynamic reconfiguration framework that allows to
autonomously substitute faulty components in an embedded system.
Description
Many cyber-physical systems have to be fault tolerant. This is typically
achieved by applying active replication, and thereby introducing explicit
redundancy (e.g., Triple Modular Redundancy). The type of fault tolerance
mechanism applied and the degree of redundancy are based on a fault hypothesis
that describes which kinds of faults are more or less likely to appear in
the system. In case a system component experiences a fault that is not covered
by the fault hypothesis, the whole system might fail.
In addition to explicit redundancy, many systems comprise implicitly redundant information,
that is defined by the interconnection of distinct components in the system
(e.g., the speed of a car has a certain relation to the rotational speed of the
car's engine). With an appropriate model, this implicit redundancy can be exploited
to dynamically reconfigure the system and thereby minimize the probability of a
system failure even in case of unexpected faults.
It is the target of this project to implement a framework for dynamic reconfiguration
based on the Robot Operating System (ROS), which is a widely used platform for robotic
applications. The framework should use an existing algorithm that is working on an
ontology-based system model to find redundant information paths in the system, and
reconfigure ROS software modules to substitute erroneous system components.
The tasks to be accomplished involve:
- design of appropriate software module interfaces
- automatic generation of ROS software modules based on the found redundant information path
- implementation of reconfiguration mechanism
Required Skills
In order to develop the framework, experience with C++ and Python is highly appreciated.
Supervisors
Univ.Ass. Dipl.-Ing. Oliver HÖFTBERGER
(main responsibility)
E-Mail:
Types
Diplomarbeit