Knowledge-driven opto-acoustic scene analysis based on an object-oriented world modelling approach for humanoid robots

The perception of the environment using opto-acoustic scene analysis techniques is an extremely important and challenging task for humanoid robots. Thereby, the acquisition of visual and acoustic information provides the basis of the perception and enables the robot to interact with humans and to assist in everyday tasks. Another step is the exploration that allows the robot to actively gather knowledge about the environment. Hereby, the object-oriented world modeling guarantees the maintenance of permanent situation awareness, which is important for opto-acoustic scene analysis, cognition and planning. Taking these considerations into account, this paper presents a concept for the knowledge-driven opto-acoustic scene analysis based on an object-oriented modeling approach. The focus of the paper is put on both, formal conditions and a practical realization of an integrated system.