The Zurich University of Applied Sciences Winterthur regularly takes part in the annual Smartrob EPFL Robot Contest, a contest where autonomous robots are programmed to solve different tasks. One of the biggest problems in autonomous robot control is autonomous robot navigation. It consists of the robot having to compute a best (shortest) path based on a map and data from sensors. No matter what specific task has to be solved, autonomous navigation is always part of it. This problem was already solved for small rooms and a limited number of predefined obstacles in a project lanced by Halm Reusser & Ranko Veselinovic. The goal of the present diploma thesis is to solve the same problem for larger rooms, like a corridor, with an unlimited number of unknown obstacles. For this purpose, an algorithm named A* Algorithm was used, or rather one of its derivations named D* Lite Algorithm, which is very flexible and powerful. The implementation uses a dynamically resizable map and is not bound to specific sensors. The outcome of this diploma thesis is a robot control that allows the robot to drive to a given target in known or unknown terrain. This robot control is based on a reusable navigation-module. The robot uses sensors to detect obstacles and landmarks. Landmarks can be predefined in the map so that the robot can adjust his position and orientation. The navigation-module can be adapted to different terrains and mechanical constructions of the robot itself, including it's sensors - an ability which renders it highly flexible and reusable. From this project we conclude that the task of finding a shortest path around obstacles is complicated, even in quite simple terrain. But the A* Algorithm and it's derivations solve this problem very well without using too much computation time, such that they can be used for smaller systems as well.