NEUROCOGNITIVE ALGORITHMS FOR MANAGING MULTI-AGENT ROBOTICS SYSTEM FOR AGRICULTURAL PURPOSES
Kantemir Ch. Bzhikhatlov, Пшенокова Инна Ауесовна, A.R. Makoev · Известия Южного федерального университета. Технические науки · 2024
The main goals of the introduction of robots into agriculture are to increase efficiency and performance,fulfilling labor -intensive and dangerous tasks and solving the issue of lack of labor. Technologicalachievements in the field of detection and management, as well as machine learning allowed autonomousrobots to perform more agricultural tasks. Such tasks vary at all stages of cultivation: from preparation ofland and sowing to monitoring and harvesting. Some agricultural robots are already available, and it isexpected that in the coming years there will be even more, since technologies for processing big data, machine vision and easy capture are becoming more accurate. Currently, the introduction of several interactingrobots in the field is becoming increasingly relevant, since it has good prospects in reducingproduction costs and increasing operating efficiency. The purpose of this study is to develop an intellectualsystem for managing a mobile robot group based on multi -agent neurocognitive architectures. The taskof the study is to develop neurocognitive algorithms for controlling the multi -agent robotics system ofagricultural purposes. The work describes a multi -agent robotics complex for active plant protectionwithin the framework of the Smart Field system. The concept of the management system of the group ofmobile robots based on modeling multi -group neurocognitive architectures is presented. To ensure thework of the multi -agent heterogeneous group of autonomous robots, the use of a neurocognitive controlmodel with the implementation of individual intellectual agents is proposed on each individual robot andat the bases of service or servers. At the same time, given the implementation of recursing in architectureitself, the task of scaling such a management system is noticeably simplified. The use of sensors and effectorsto ensure the exchange of knowledge between robots and decision -making centers allows minimizingthe load on the communication system and ensure a reserve of failure tolerance of the management system.The results obtained can be used to develop universal control systems and simplification for variousgroups of autonomous robots.