REVISITING PL UCKER COORDINATES IN VISION-BASED CONTROL

Abstract Plucker Coordinates are largely used in kinematics. What is perhaps much less known is the fact that such a representation of lines exhibits very interesting properties in some classes of control problems. In this paper we start from classical Plucker Coordinates and show that, by adding a supplementary normality constraint, we obtain a representa-tion which is very ecient for controlling a system from images of lines obtained through a perspective projection. Furthermore, the physical interpretation of the approach is very easy. Having derivated the related equations of motion, we then design a control scheme the stability of which can be proved. Some simulation and experimental results illus-trate the approach