The article discusses and solves problems associated with the development of one of the approaches to constructing a model for representing the procedural knowledge of an intelligent robot in the form of a given set of behavioral frame-microprograms. In general, frame microprograms represent behavioral skills associated with solving the corresponding subtasks, without reference to a specific subject area, providing the robot with the ability to automatically generate plans for goal-directed activities related to solving complex problems in various environmental conditions. In other words, frame-microprograms of behavior are the “building blocks” on the basis of which an intelligent robot can construct complex plans for purposeful activity in various operating conditions. In development of this model for representing a robot's procedural knowledge, a structure of a frame-microprogram is proposed, as a result of the decomposition of which it is possible to obtain modules that define relatively simple behavioral skills. This approach allows the robot to organize either a more flexible process of planning purposeful activities in underdetermined problem environments, or to provide the intelligent robot with the ability to automatically construct typical elements of procedural knowledge representation that are missing to achieve a given goal. The choice of a particular one of these two methods of using the modules obtained as a result of splitting the behavioral frame- microprograms depends on the complexity of the task being solved by the intelligent robot and the dimension of the semantic network describing the current situation of the problem environment. The conditions have been determined to ensure the identification and selection of effective knowledge representation modules in the process of planning targeted activities and the automatic construction of behavioral frame- microprograms that are missing for the conclusion of decisions. It is shown that the proposed approach to constructing a model for representing procedural knowledge and tools for processing it in the process of planning purposeful activities makes it possible to increase the cognitive capabilities of intelligent robots and, on this basis, provide solutions to complex problems in underdetermined problem environments.
DOI: 10.37220/MIT.2024.64.2.019
Download the journal (PDF) from the official website (in Russian): http://morintex.ru/wp-content/files_mf/1717093978MIT2PART12014.pdf
Download the article (PDF) from eLibrary (in Russian, registration required): https://www.elibrary.ru/item.asp?id=68557177
Vladimir B. Melekhin, Mikhail V. Khachumov. Automatic construction of procedural knowledge representation elements by an underwater intelligent robot under conditions of uncertainty // Marine intellectual technologies, № 2, part 1, 2024. – Pp. 155–162.