TASK TRANSLATION AND INTEGRATION SPECIFICATION IN INTELLIGENT MACHINES

Intelligent machines are defined to be hierarchically intelligent cont rol systems composed of three levels: the organization level, the coor dination level, and the execution level. This paper presents an analyt ical model for the coordination level of intelligent machines, which, together with the established mathematical formulation for the organiz ation level and the well developed control theory for the execution le vel, completes the first step toward a mathematical theory for intelli gent machines. The framework of the coordination level is a tree struc ture consisting of a dispatcher and a number of coordinators. A new ty pe of transducers, Petri net transducers (PNTs), has been introduced t o serve as the basic module in our analytical model. PNTs provide a fo rmal description for the individual processes within the dispatcher an d coordinators. The concurrence and conflict among these processes can be represented by PNTs conveniently. Coordination structures are intr oduced as a formalism for the specification of integration in the coor dination level. The task precedence relationship in the coordination p rocess is presented by the Petri nets derived from the coordination st ructures. These Petri nets also provide us a formal approach of using the concepts and analysis methods in the Petri net theory to investiga te the properties of the coordination structures. A case study of mode ling an intelligent assembly robotic system has been conducted for the purpose of illustration.