Abstract
The present bachelor thesis presents the necessary methods for an exact selflocalization by using an Inertial Mesurment Unit (IMU) and the odometry of an autonomous lawn mower. This self-localization shall be used in later work together with a localization of a particle filter. The required standard models [12] for the individual sensor systems were examined and the required parameters determined. Measurements were taken with the autonomous lawn mower to develop a Kalman filter [15] based on the data obtained. With the help of the Kalman filter and a controller which was designed in this thesis, such a self-localization could be realized. The results show that the autonomous lawnmower can locate itself under simulated conditions with a higher accuracy than with a pure localization via odometry. However, final measurements show that disturbances of the IMU occur within the autonomous lawn mower, so that a final overall behaviour cannot be implemented with the current architecture of the autonomous lawn mower.