Hybrid neural networks in cyber physical system interface control systems

Sergei Sokolov,Nikolay Glebov,Anatoliy Nyrkov,Sergei Chernyi,Anton Zhilenkov

doi:10.11591/eei.v9i3.1293

Abstract

The calculation and results of simulation of the magnetic control system for the spacecraft momentum are presented in the paper. The simulation includes an assessment of the reliability of calculating the Earth's magnetic field parameters, as well as an assessment of the quality of object stabilization by resetting the total momentum with the aid of the system under review. The outcome of a comparative analysis of resource efficiency and energy efficiency are demonstrated in the implementation of the proposed hardware models of controllers on FPGA. The strengths and weaknesses of the programming models are shown. The developed models will allow to be modified and perform more complex operations in the future.

Highlights

The technology of the artificial neural network is designed to provide the machines of the future with complete autonomy, i.e. autonomous control and decision-making, relying on the previous experience and analysis without direct human involvement
The control principle is based on the standard form of the transfer function of the propulsion system, using the transfer functions and linearization of any nonlinearities inherent in thermodynamic, physicochemical processes, for example, and combustion
In designing the control systems for biomechanics devices, the most difficult task is the processing of pulses from the human body, recorded by means of such techniques as electromyography (EMG) and electroencephalography (EEG), and used to control, for example, prostheses, orthoses for rehabilitation purposes, exoskeletons etc

Summary

INTRODUCTION

The technology of the artificial neural network is designed to provide the machines of the future with complete autonomy, i.e. autonomous control and decision-making, relying on the previous experience and analysis without direct human involvement. Artificial neural networks, being a model of biological neural networks, are applied to solve problems, the algorithm for solving which is unknown [1,2,3]. The standard PI (D) control law forms the main linear module of the vast majority of modern electronic diesel engine controllers. The control principle is based on the standard form of the transfer function of the propulsion system, using the transfer functions and linearization of any nonlinearities inherent in thermodynamic, physicochemical processes, for example, and combustion. In designing the control systems for biomechanics devices, the most difficult task is the processing of pulses from the human body, recorded by means of such techniques as electromyography (EMG) and electroencephalography (EEG), and used to control, for example, prostheses, orthoses for rehabilitation purposes, exoskeletons etc.

ADVANTAGES OF THE PROPOSED APPROACH

VIRTUAL MODELING

CONCLUSIONS