Obstacle recognition in the workspace of an industrial robot by a 3d-PMD-Camera

Presently security fences are often installed to prevent the entering of humans in the robot workcell. The installation of these fences is very expensive, inflexible and it isn't place-saving. Furthermore they avoid the interaction between humans and robots, but such interactions are preferable for many industrial applications. In this work a new approach is described, which based on the use of a special 3d-PMD-Camera to observe the robot workcell. This camera facilitates the detection of both, humans and obstacles, in the workcell. So it is possible for the robot to react on changes at his environment. Such reactions can be the adjustment of the robot velocity or stopping the execution of the movement, up to planning a new collision free trajectory based on the PMD-Camera generated 3d-data.