Localization of object edges in arbitrary spatial positions based on ultrasonic data

In mobile robotics, ultrasonic sensors became standard devices for collision avoiding. Moreover, their applicability for map building and navigation has exploited in recent years. Although three-dimensional (3-D) map building methods arose in recent years, most methods are concentrated on two-dimensional data representation. In this work, a new method for 3-D localization of object edges in arbitrary spatial positions and orientation is presented. With the help of a 3-D map of geometric modeled object edges, a high flexibility in the application of ongoing measurement values for the continuous correction of the robots´ odometric drift can be achieved. The main problems with the practical application of this method are the statistical and distance-dependant behavior of conventional ultrasonic sensors, especially at the border of the sound lobe where the accuracy of the sensors normally decreases. Therefore, ways to improve the sensor accuracy for modeling and position measurement by enhancing sensor performance are shown. Finally, this paper includes an outlook on specialized sensors, applying sophisticated signal processing methods, especially suited for the localization method presented in this work.