Electronic Device Control Method Based on Improved Neural Network Algorithm


  • Meirong Gao


Smart home appliances are one of the development directions of home appliances. Using
artificial neural network or fuzzy artificial neural network to realize intelligent control of
electrical appliances has the characteristics of powerful functions and strong
adaptability. The control of household appliances is relatively simple, but the artificial
neural networks currently used in industry are implemented using large-scale field
programmable gate arrays (FPGAs), which is not cost-effective for general electrical
applications. The purpose of this research is to study the artificial neural network control
of electrical equipment based on simple analog circuits. Has the following
characteristics: to meet the basic requirements of automatic control of electrical
equipment; simple structure, low cost, strong economic and practical; simple learning
and training, easy to operate, suitable for large-scale production; with general
characteristics, that is, different learning and training can Different application needs.
This article uses PSpice circuit simulation software to model, simulate and optimize the
circuit. In this paper, PSpice simulation software is used to model, simulate and optimize
the nonlinear function generator circuit, adder circuit and analog multiplier circuit, and
finally obtain the expected results. PSpice circuit simulation software is convenient and
fast. Using simulation results to guide the experiment, you can achieve twice the result
with half the effort. This article designs a load cell that can convert different weights into
electrical signals, then normalize them, and then amplify the signals as the input to the
ANN. Using the error analysis of the ANN overall circuit, the calculated maximum error
is only 1.17%. In the end, this article draws the results of a simple artificial neural
network circuit design scheme that is simple in structure, low in cost, and common for
learning and training programs, which is convenient for manufacturers to mass produce.