Automatische Bahnplanung und Hindernisumfahrung für ein autonom navigierendes Fahrzeug

Automatic path planning is an important prerequisite for various applications in robotics, such as autonomous vehicles or automatic guided vehicle systems. In this thesis, different approaches for path planning are presented. The methods are compared and evaluated, especially considering their ability to handle differential constraints. Then, the concrete implementation of a sampling-based approach using motion primitives is explained. This algorithm is finally expanded to allow dynamic replanning in reaction of unexpected environmental changes. In numerous simulations, the implemented method proves its ability to solve path planning problems under a variety of conditions in a relatively short amount of time. It is only in extremely narrow spaces and highly dynamic environments when the approach fails to generate a solution.