Aarhus Universitets segl

Autonomous Intelligent Systems

Theme focus

Autonomous systems have to be able to perceive their environment and adapt their own behaviour in order to continuously operate towards their goals. Autonomous Intelligent systems require more than simple self-adaptation approaches but a deeper understanding of themselves and their environment. As such, we consider both, pure software systems as well as cyber-physical systems, controlled by autonomous software agents, alike. Operating in a dynamic environment and being subject to rapidly unfolding situations, autonomous intelligent systems have to respond rapidly to potentially unknown circumstances.

To accomplish this, such systems have to bring together several abilities:

  • to learn about themselves and their environment,
  • to model the states they encounter consistently and identify similarities comparable situations,
  • to adjust their behaviour in accordance with their surroundings,
  • to verify their actions and states,
  • to make profound decisions based on their observations, their verifications, and their previous experiences.

Furthermore, to tackle previously unknown situations in the presence of other systems, Autonomous Intelligent Systems will have to

  • transfer previous experiences into new situations,
  • learn about others, integrate themselves, and establish autonomous collaboration where feasible without being exploited or exploiting others,
  • establish and utilise trust to other systems in their environment to achieve goals more efficiently,
  • halt and reflect on their experiences, or
  • respond quickly based on best approximate solutions, and finally
  • be aware of their goals and select them appropriately during runtime.

The Autonomous Intelligent Systems theme contributes to the individual research in these areas taking inspiration from biology, physics, social science, or psychology. Furthermore, we also develop systems combining existing approaches and techniques to identify shortcomings and unearth future directions of research. For this the theme closely collaborates with other research themes at Aarhus University and various research teams around the world.

Teaching activities

Developing Autonomous Intelligent Systems requires many different competences. Several courses are already taught at AU and should be considered by students desiring working and doing research in this area.

  • (Advanced) Signal Processing
  • Computer Vision and Machine Learning
  • Modelling and Verification
  • Control theory