The human brain normally finds it easy to recognize objects reliably. For example, it can usually identify a car reliably, regardless of the viewing angle and distance. This is a difficult task for computers - they have to be painstakingly taught to recognize objects. This is exactly what Oliver Klöckner (19) from Chemnitz did in his research work. His aim was to record the positions and angles of two joints of a special pendulum. To do this, he attached distinctive geometric figures to the joints. When a camera films the movement of the pendulum, his self-written software reliably recognizes the desired spatial data of the joints.