Aspects of cognitive understanding of the environment by vision-based semantic mapping
Quirin Mühlbauer · 2010
The role of mobile robots in our daily lives has been increasing rapidly during the last decades. Essential elements of such systems are sophisticated perceptual and cognitive skills, extraction and storage of knowledge, and methods for action planning. The main contributions presented in this thesis are vision-based methods for object detection and mapping, a cognitive system for knowledge representation and decision making based on semantic maps, and the integration in an autonomous mobile robot. By advancing the perceptual and cognitive skills of mobile robots, the proposed methods support navigation in real urban environments. Simulations and experiments have been conducted to verify the presented cognitive abilities.