ABSTRACT
Geometry, in its quantitative connotation, is a favored means of analysis to improve the technology of robots in the assembly process. This is because the objects are usually supposed to be rigid and the corresponding measures invari ant, but other implicit assumptions are hidden in this sort of description of the robot’s world, which limit what we can expect from a system. Later, we will discuss these assumptions and their consequences. Although it is clear that the geometric view is not sufficient, perhaps even unsuitable for advanced applica tions in unstructured environments, here we restrict our attention to those prob lems that can be approached by means of accurate geometric information.