SEPIAS - Self Properties In Autonomous Systems

The Self Properties in Automomous Systems project (SEPIAS) focusses on the development process of mobile systems which comprises hardware and software. The goal is to provide tools and techniques for enabling the mobile systems to react to environmental changes and internal faults in an intelligent way such that they can still fulfil their tasks. Because of limited possibilities for human intervention and even interaction during operation autonomous systems need to solve the problems and unexpected situations by their own. As a consequence, the autonomous system has to have knowledge about the surrounding environment, its task, and itself. Moreover, the system has to be able to reason about its state and if necessary to draw conclusions which even have an effect on its structure and behaviour. For the capabilities of reflection and self-healing capabilities the required knowledge has to be accessible by the system development of hardware and software of a system for a specific task. Although malfunctions are considered partially within the process, the resulting system has no explicit knowledge about itself. To overcome this situation, there has to be a knowledge acquisition phase during the development process which integrates smoothly within a traditional development process. For the purpose of knowledge representation and reasoning this project uses the techniques that are used in model-based reasoning. They have to be adapted and smoothly integrated. An additional requirement is the available computational power of autonomous systems since reflection has to be performed during operation. Hence, the algorithms and knowledge representation have to be adapted to perform well under given computational restrictions.