Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Self-adaptive control method for embedded systems O.V. NEPOMNYASHCHIY, N.Y. SIROTININA, N.A. MAMBETALIEV, V.V. GOREVA, H.A. LATYSHONOK This paper describes a self-adaptive control method for embedded systems. The method is based on a self-learning intelligent agent. A decision-making system of the intelligent agent uses procedural reasoning system, supplemented with building the new scenario of achieving goals through genetic algorithm. A proposed software-hardware implementation of the genetic algorithm provides both maintaining its flexibility and reducing scenario search time. The described method has been modeled and tested on the traveling salesman problem. Simulation results allow us to define the most computing-intensive procedures of the control algorithm. The architecture of hardware and software complex for implementing the proposed method is developed on the VLSI “System-on-a-Chip”. Through combining hardware and software implementation of the algorithm of searching and building a new scenario, the described solution provides reducing the computational complexity of control processes and extending the functionality of systems. Keywords: Embedded systems, intelligent agent, decision support systems, adaptive control, System-on-a-Chip, genetic algorithm, FPGA. Contacts: E-mail: 2955005@gmail.com Pp. 16-21.