Intelligent Autonomous Systems Group (ISLA), University of Amsterdam
The Intelligent Autonomous Systems Group is one of the three groups that make up the University of Amsterdam’s Intelligent Systems Lab Amsterdam, or ISLA. Here, the group carries out teaching and research in the field of multi-agent systems and cooperative robotics.
The focus of the group is real world multi-agent systems and formal techniques for describing system and agent interactions. The people behind IAS study methodologies to develop intelligent autonomous systems that perceive their environment through the use of sensors and utilize that info to produce intelligent, goal-directed activities.
Part of their work comprises formalization, generalization and learning of goal-directed behavior in an autonomous system. Specific projects include perception, geometric principles, multi-agent systems, and learning.
IAS’s research focuses on three specific areas. The first is uncertainty representation and shared world models. Here, using Markov Decision Processes, they are able to see how different robots can use sufficient levels of shared knowledge in order to do a task successfully.
Another is role determination and team formation, wherein a team sees a role in an intuitive and flexible way. In this role, they also extend the framework of the Multi-agent Markov Decision Processes (MMDP) in order to integrate the notion of a role, and show how they can use it in a real robot soccer game.
The third part is the apt software architectures for real-time multi-agent systems. Here, allowing a large number of processes to communicate asynchronously using virtual shared data space is made possible.
One of the projects that IAS has successfully developed was done in collaboration with the Vrije Universiteit Amsterdam. It was called the “A Framework” and was used to maintain a shared model in a dynamic environment between differentiated embedded systems, making interaction with humans possible. The project was expanded for two years and has involved people outside the school.