This paper presents an approach to fast mapping of obstacles from a workspace into a configuration space, based on defining some specific points in the workspace as fundamental obstacles. In order to achieve this objective, we propose efficient algorithms for mapping complex Cartesian obstacles by determining their critical points. The computational time required for mapping a two-dimensional obstacle is approximately 5 ms with a 33 Mhz 80486 CPU. Obstacle mapping in this paper is based on a CAD model of an environment that provide global information. The proposed mapping procedure and algorithms are suitable for industrial manipulators with a symmetric workspace occupied by polygonal obstacles.