Abstract. Whenever mobile robots are used in real world applications, the ability to learn an accurate model of the environment and to localize itself based on such a model are important prerequisites for reliable operation. Whereas these problems have been successfully solved in the past for most indoor tasks, in which the robot is assumed to operate on a flat surface, such approaches are likely to fail in combined indoor and outdoor environments in which the three-dimensional structure of the world needs to be considered. In this paper, we consider the problem of localizing a vehicle that operates in 3D indoor as well as outdoor settings. Our approach is entirely probabilistic and does not rely on GPS information. It is based on so-called multi-level surface maps which are an extension of the well-known elevation maps. In addition to that, we present a technique that allows the robot to actively explore the environment. This algorithm applies a decision-theoretic approach and considers the uncertainty in the model to determine the next action to be executed. In practical experiments, we illustrate the properties as well as advantages of our approach compared to other techniques.

@InProceedings{pfaff07irosws,
  author = 	 {Patrick Pfaff and Rainer K\"{u}mmerle and 
                  Dominik Joho and Cyrill Stachniss and 
                  Rudolph Triebel and Wolfram Burgard},
  title = 	 {Navigation in Combined Outdoor and Indoor 
                  Environments using Multi-Level Surface Maps},
  booktitle =    {Proc. of the Workshop on Safe Navigation in 
                  Open and Dynamic Environments at the {IEEE} Int. Conf. 
                  on Intelligent Robots and Systems {(IROS)}},
  year = 	 {2007},
  month =        oct,
  address = 	 {San Diego, CA, USA}
}