Abstract:
The paper considers the task of control a mobile agent in a dynamic unknown environment. To solve this problem, it is proposed to use a neuro-fuzzy controller. To adjust the parameters of the controller, a genetic algorithm is used. The paper also proposes a modification of the genetic algorithm that allows to configure not only the parameters of the membership functions of linguistic variables, but also their type. This allows to fine-tune the controller and, as a result, improve the quality of the system. The effectiveness of the proposed approach was confirmed by computer simulation in the developed virtual environment. The resulting controller was used in a physical experiment with a mobile agent implemented on the basis of Arduino Uno, in an environment with dynamic and static obstacles.
Fulltext:
PDF file (642 kB)