The goal of this thesis is to allow for mobile exploration of an autonomous robot in a human-made indoor environment by creating and implementing a robotic system that can build up and use a map of the surroundings in a fully autonomous way. In such a setting the most dominant parts are built out of planar surfaces such as walls, tables, furniture, etc., therefore planes are a compact and easy way to describe the environment. A plane-based mapping algorithm is used to create a map of the current environment and enable mobile navigation. The used topological mapping approach has the advantage that it is not only a tight representation of the situation but also needs only minimal memory requirements, as planes can be described with a small amount of data. Further on an object detection is implemented to expand this initially purely topological map to a more descriptive representation of the surroundings including objects and information about these. The system allows for autonomous explorative interactions with these objects to automatically acquire various object features, such as their degrees of freedom, and expand the knowledge about the world. The whole system is integrated into the Machine Learning and Robotic code of the University of Stuttgart and is tailored to work on a PR2 robot using a Microsoft Kinect camera for visual perception.