Heinrich Mellmann (HU Berlin)

2010/11/12, 16:00

HU Berlin, at the BMS Seminar Room (RUD 25, 1.023)

Whatever we do, wherever we go, most of the time we know “where” we are—in Berlin, at home, in front of the fridge, etc. For a successful interaction with the environment (e.g., switching on the TV set) the knowledge of one's own position relative to the object is essential. Thus, an artificial being, like a robot, which should perform similar to a human also has to know its environment and especially its position in it. In this talk we introduce some mathematical groundings for Bayesian modeling and discuss, in particular, the Monte-Carlo particle filter which can be used to model the position of a robot in a dynamic world. We will also illustrate the gray theory with interesting videos and examples from robot soccer.