ROBOT PROGRAMMING SYSTEM FOR ASSEMBLY - CONCEPTUAL GRAPH-BASED APPROACH

This work develops a method for robot program synthesis. Currently, th e programming task is one of the major hurdles of robotic application. Progress towards automatic synthesis of robot programs will ease indu strial robot application. The proposed system provides a means towards automated (guided by knowledge) conversion of a user's request, expre ssed in natural language, to the appropriate conceptual model of the r equired task. This model incorporates the information necessary for un derstanding, planning, and sensory-guided performance of the required robotic task. First, we state the problem of robotic assembly and reco gnize its hierarchic structure as the structure of a system that build s a predesigned assembly. Next, we present and analyze the requirement s of the robotic assembly domain. This analysis enables us to draw con clusions concerning the most suitable methodology for the development of a support system for assembly program synthesis and interpretation. It is the conceptual graph-based approach. Then we present the algori thm of the proposed conceptual graph-based system and show how the sys tem synthesizes robotic assembly operations, such as valid assembly se quences and sequences of special treatments for the assembled componen ts (including the determination of the required resources). Finally, a case study illustrates the approach developed here on a large family of multi-axisymmetric components. We also present illustrative example s of working sessions with the current implementation of the system.