Abstract:
An approach to solving the problem of cognitive control of robotic manipulators is proposed on the basis of converting information coming from a technical vision system into a mathematical model of a cognitive digital automaton. The cognitiveness of the latter one is determined by the possibility of learning the trajectories of the manipulator's movement and methods of grasping an object, as well as predicting (generating) solutions in situations not provided for in the learning process. The scientific novelty of the method lies in the formalization of the concept of cognitive control of robotic manipulators.
Keywords:cognitive control method, cognitive automata, neural networks, mathematical model, Petri nets, equation of states, logic, learning, decision prediction, robotic manipulator.
Fulltext:
PDF file (686 kB)