Autonomous systems solve complex problems, they learn, make decisions independently and react to variable or unknown situations without human intervention. For reliable and efficient use in the Smart Factory, however, it is crucial that autonomous systems are able to maintain their functionality even in the event of a malfunction by initiating suitable self-healing operations. This is where previous approaches to solutions reach their limits.
In the KI4AS project, artificial immune systems are researched using combined methods of artificial intelligence, machine learning and biologically inspired algorithms. It aims at the implementation of self-healing properties to ensure the health and functionality of autonomous systems.
Analogous to biological immune systems, self-healing systems observe and analyse themselves and search for abnormal signals and data. When a functional error occurs, they use part of their resources to make an independent diagnosis of the causes and to plan and implement behavioural adjustments to restore functions.
The use of artificial immune systems thus ensures the safety of the systems at all times and increases the overall reliability of the system. At the same time, life cycle costs are reduced and the ecological sustainability of the system is increased.