Модель навигации робототехнического комплекса в многокомпонентной информационной среде

Abstract

Purpose: In the design of modern marine robotic complexes, various problems arise: analyzing the informational needs of autonomous agents; studying and modeling the information structures and processes associated with the search, acquisition, accumulation and processing of the information by artificial cognitive agents; the development of estimates, opinions and knowledge by artificial cognitive agents on the basis of the information they process. To solve these problems, the authors propose a general model of a navigational robotic complex based on agent-oriented methodology where the systems and their components are interpreted as cognitive agents. Results: The main features of developing robotic complexes using agent-oriented approach were studied, including the architectural features and intelligent agent functioning. A feature of artificial cognitive agents was described which distinguishes them from other intelligent systems – getting information from three sources: directly from a human, from a sensory system and from its own database/knowledge base. A general scheme was developed of a combined navigation for a cognitive agent, which provides its functioning in two different modes: "rough" navigation associated with the movement of the agent to the target on the base of the granular information, and "accurate" navigation based on the current numerical data about the environment obtained from the sensor system. A hierarchical dual control system was formed where the classical control scheme is complemented by a fuzzy control subsystem. Practical relevance: The proposed navigation model based on the agent-oriented approach can be used in the design of marine robotic complexes for both civilian and military purposes, which will greatly reduce the complexity of the design as a whole.