Robots are built to assist people. Up to now the circle of people benefiting is still rather small because operating a robot is a very difficult task. To enable untrained people to operate a robot, an intuitive user interface has to be provided. Besides enlarging the number of people which are actually able to operate the robot, the operators are less burdened by controlling the robot and therefore are able to concentrate better on the task they actually have to perform. This paper presents a method of applying the maybe most intuitive modality of control: the operator can give commands to the six-legged walking robot LAURON by pushing or pulling its body. To achieve this the robot's 18 joint torques are calculated based on the motor currents. These are then transformed and merged in a 6D force vector corresponding to LAURON's CoG. This vector is finally fed into a classifier which identifies the user's intention out of a set of possible commands and triggers corresponding behaviors.
展开▼
