|
|
Christoph UrbanczikUniversity of Kaiserslautern Informatics
Faculty
|
Diploma Thesis: "Sensor based Collision Detection and Path Smoothing in C-Space" (Presentation Slides) (Part of SIMERO Project)
Demonstration Video (AVI MPEG4,2.2 MB)
This work deales with the detection of obstacles in a working space of an industrial robot with the help of several digital cameras. Therefor a volume-based algorithm is introduced, which should .speed up an ordinary collision test. The basic idea is, that the necessary collision-tests should be reduced to a minimum. This means, that only the volume, which is crossed by the robot during its movement, should be tested for collision.
The other part deals with the smoothing of a randomized-road-map-plan. The goal is to get a smooth and short path, which provides a collision-free path between two configurations. Therefor the algorithm tries to reduce the length of the path-segments provided by the planer and avoid sharp changes of direction at the roadpoints.
Both algorithm are designed for the use with a randomized-road-map-planer. They take advantage of its special feature, that the generated path always consists of linear segments in c-space.
Last update: