Functional description of various inference modes in a hierarchical model: application to robotics

Human specialists deal with growing complexity of robots design, control and maintenance by running experienced nonrigorous reasoning on their own description of the system (mixing detailed and global information). Representation of such an approach is looked for in this paper: nonrigorous types of reasoning and hierarchical modelling of complex systems are mixed together with a proposed approach. It is applied to diagnosis of a complex robotics system.<
>