On Environmental Model-Based Visual Perception for Humanoids

摘要

In this article an autonomous visual perception framework for humanoids is presented. This model-based framework exploits the available knowledge and the context acquired during global localization in order to overcome the limitations of pure data-driven approaches. The reasoning for perception and the properceptive components are the key elements to solve complex visual assertion queries with a proficient performance. Experimental evaluation with the humanoid robot ARMAR-IIIa is presented.