Welcome
The Robot Learning Group does research in the field of autonomous adaptive robotics. An autonomous
robot is a machine that operates in a partially unknown, changing, and unpredictable environment. These characteristics
make it difficult to employ the traditional engineering approach based on problem analysis, modelization, and design.
By equipping our robots with adaptation mechanisms, such as learning by experience, we create systems capable of operating
in real environmnents.
Our aim is that of understanding the basic and simple principles of adaptation that can produce complex, robust, and reliable
behaviors for autonomous mobile robots. The learning mechanisms and control architectures that we develop are platform-independent,
which means that they can be applied to a variety of mobile robots.
An important point of our research is biological inspiration. Autonomous robots and biological organisms face similar
challenges in that they both must survive in a partially unknown environment. Since biological organisms display interesting
characteristics of adaptation, robustness, and self-sufficiency, understanding the computational principles of biological
self-organization is a fruitful enterprise also for engineering robust artificial controllers. At the same time, real robots are
an optimal tool for testing theories and models of biological adaptation.
We pay considerable attention to engineering issues, such as reliability, feasibility, and utility. We are therefore
concerned with the issues of generalization, amount of human design, and cost-to-benefit ratio. One of our goals is that of
building solutions that can be easily transferred to real world applications.
Our approach is highly interdisciplinary as well as the background of our members. As such, our group benefits from continuous
and intense exchanges with several other members of the Laboratory
of MicroComputing, such as people of the Mobile Robots
Group, the Robot Biovision Group,
the Collective Robotics Group, and
the Neural Computation Group.