Hidden Markov Model Based Visual Perception Filtering in Robotic Soccer

Autonomous robots can initiate their mission plans only after gathering sufficient information aboutthe environment. Therefore reliable perception information plays a major role in the overall success of anautonomous robot. The Hidden Markov Model based post-perception filtering module proposed in this paper aimsto identify and remove spurious perception information in a given perception sequence using the generic metaposedefinition. This method allows representing uncertainty in more abstract terms compared to the commonphysical representations. Our experiments with the four legged AIBO robot indicated that the proposed moduleimproved perception and localization performance significantly