Organization of Interaction of Autonomous Intelligent Robots in the Process of Joint Solution of Complex Subtasks

The article considers a solution to one of the complex problems associated with the organization of collective behavior of autonomous intelligent robots, when in order to perform a complex subtask on a given section of the problem environment, obtained by dividing the general task into autonomous subtasks, it is necessary to involve several robots for joint purposeful activity. Various in capabilities, purpose and structure elements of knowledge representation of intelligent robots have been developed, regardless of a specific subject area, providing the ability to organize the search for solutions to subtasks of varying complexity under uncertainty. In particular, planning the joint purposeful activity of several robots based on frame micro programs of behavior that determine the solution of elementary typical subtasks allows to significantly reducing the search space by defining a number of effective actions for various intelligent agents at each step of the search for a solution to complex subtasks. In turn, planning of joint purposeful activity of intelligent robots based on frames of relations and actions, as well as micro programs of behavior associated with the transfer of certain objects from the current to a given state when it is necessary to eliminate individual differences between the initial and target situation of the problem environment, ensures flexibility in finding solutions to subtasks of varying complexity. This is achieved due to the fact that the proposed decision-making tools allow groups of robots, differing in number, to effectively plan joint purposeful activity associated with solving the subtasks assigned to them by rationally combining elements of the knowledge representation model with different purposes. The main operations performed in the process of decision-making are the operations of determining fuzzy nested equality and fuzzy equality between different semantic networks. In general, the developed model of knowledge representation and processing allows creating problem solvers that allow organizing joint purposeful activity of intelligent robots in the process of solving problems and subtasks of varying complexity in a priori undescribed problem environments.