Dedicated mechanisms for the attention system in humanoid robots

A humanoid robot must be able to cooperate efficiently and safely with humans in an unconstrained environment. Furthermore, it should not only present no danger to humans itself, but also be able to detect dangerous situations for the user. To restrict the processing only to salient objects and events, attention systems modeling the human visual system have been proposed. However, they are computationally demanding and the capabilities of digital signal processing systems are still rather limited compared with those of humans. Therefore, we propose that simple dedicated mechanisms should be introduced to identify important or even dangerous events. To this end, efficient detectors for objects falling down, objects moving at high velocity as well as human faces are presented in this work.