Semantic Environment Perception, Localization and Mapping

Abstract

The sensory acquisition of the environment is the most important task of mobile robotics, as it is the foundation for any ability that the robot shall have, later on. Sophisticated tasks often require an environment model for path planning,obstacle avoidance and many more. Furthermore, the robot needs to know where it is located within the environment to build-up, complement and update the model. Thus, besides environment perception, localization belongs to the most important tasks of mobile robot systems. Most approaches towards self-localization and mapping are very specific, either to one sensor type, or a strictly predefined set of sensors, prohibiting the use of the provided techniques on many different mobile systems (robots, cars or other moving platforms equipped with sensors). We present a general approach supporting the use of arbitrary numbers and types of sensors simultaneously. This allows to operate with a large variety of already existing systems without changing the hardware setup. Furthermore, the semantic environment model, generated by our solution, can directly be used for sophisticated and automated environment analyses.