This paper proposes a qualitative representation for robot kinematics in order to close the gap, raised by the perception-action problem, with a focus on intelligent connection of qualitative states to their corresponding numeric data in a robotic system. First, qualitative geometric primitives are introduced by combining a qualitative orientation component and qualitative translation component using normalisation techniques. A position in Cartesian space can be mathematically described by the scalable primitives. Secondly, qualitative robot kinematics of an n-link planar robot is derived in terms of the qualitative geometry primitives. Finally, it shows how to connect quantitativeness and qualitativeness of a robotic system. On the one hand, the integration of normalisation and domain knowledge generates normalised labels to introduce the meaningful parameters into the proposed representation. On the other hand, the normalised labels of this representation can be converted to a quantitative description using aggregation operators, whose numeric outputs can be used to generate desired trajectories based on mature interpolation techniques.
- qualitative reasoning
- qualitative modelling