Aim of the lecture
This lecture imparts the basic principles about localization, mapping and simultaneous localization and mapping (SLAM), as well as basic methods for path planning. After successful completion of the lecture, students are able to explain the principles and algorithms in SLAM and path planning. They can implement selected methods and are thus able to understand modules of available robotics packages.
Lecture content
Robot motion model. Laserscanning and landmark detection. Positioning using estimation of a similarity transform. Iterative closest point method. Bayes filter. Parametric filters and the Kalman filter. Variances and error ellipses. Extended (EKF) and multidimensional Kalman filter. Histogram- and particle filter. EKF SLAM. Rao-Blackwellized particle filter SLAM (FastSLAM). Path planning: Dijkstra and A* algorithms, potential functions, path planning in the kinematic state space. In the exercises, most of the algorithms will be programmed in the programming language Python.
30167 Hannover