Humanoid robots have become a popular research tool in recent years
and more and more research groups worldwide develop complex machines
with a human-like body plan and human-like senses. The motivation of
this research area is, first, to develop robots that are better
adapted to environments designed for humans and, second, the hope that
creating robots which resemble humans leads to a better understanding
of the human body and behavior. We believe that studying humanoid
robots will give insights how cognitive agents can understand their
surroundings by perceiving and acting in the environment as well as
how they can perform reasoning and interaction.
Several prerequisites exist to develop autonomous robots which operate
in human-populated environments. First, the robot needs to perceive
the environment with its sensors and to detect people. Second, it has
to build and maintain a model of relevant aspects of the
environment. Third, the robot should be able to interact with humans
in a natural way, i.e., using modalities humans are used to, such as
speech, gestures, and eyes-gazes. In our research, we tackle all three
problems.
Furthermore, we are interested in navigation of humanoid robots in
realistic, complex indoor environments, which contain different rooms
as well as multiple levels connected by steps and staircases.
In all our research activities, we focus on adaptive, learning
robots. We believe that key techniques to develop such agents are
probabilistic approaches that allow for dealing with uncertainty as
well as methods from machine learning.
