Artificial Intelligence in the Physical World
The Intelligent Control Systems (ICS) group aims to develop artificial intelligence for machines in the physical world. Can a machine, such as a robot or autonomous vehicle, learn new behavior from scratch? How can a team of robots efficiently coordinate their actions – what information should they exchange, and when? How to deal with limited resources such as bandwidth or energy? And: What safety guarantees can we provide about the decision and learning algorithms? These are some of the fundamental questions that arise when artificial intelligence meets the physical world – and that we address in our research.
Our research combines machine learning, systems & control theory, distributed algorithms, and robotics. The main areas of our current research are:
- Learning-based control: intersection of machine learning and control theory for improved autonomy, adaptivity and performance
- Distributed intelligence and networks: learning and control across multi-agent networks
- Resource-efficient algorithms: achieving high performance with limited resources (e.g., data, communication bandwidth, energy)
Starting from mathematical problem formulations, we develop fundamental theory, new methods, and algorithms for intelligent systems. Turning mathematical and theoretical insight into enhanced autonomy and performance of real-world physical systems is an important and driving facet of our work.
Research projects at ICS are interdisciplinary and span some or all of the above-mentioned areas. Information on some current or recent projects (including publications and videos) can be found below.