Research of Velocity Control Based on Genetic Algorithm Training RBF Neural Network

Author

Ke, Min ; Ying, Ji

Author_Institution

Dept. of Mech. Eng., Zhejiang Univ., Hangzhou, China

Volume

2

fYear

2009

fDate

Nov. 30 2009-Dec. 1 2009

Firstpage

53

Lastpage

56

Abstract

Ram velocity tracking control is an important process in injection molding control. Due to the nonlinearity of the injection system and the fluctuation of the system parameters during the process, traditional PID controller can´t satisfy the requirement of precision injection. A method of utilizing RBF neural network to adjust PID control parameters is presented, which conquers the deficiency of traditional PID controller. Genetic algorithm is used to optimize the centers and widths of hidden layer and the weights between hidden layer and output layer of RBF neural network. Gradient descent method is used to adjust the PID controller parameters. Simulations are provided to evaluate the performance of the proposed injection velocity control system.

Keywords

genetic algorithms; gradient methods; injection moulding; learning (artificial intelligence); nonlinear control systems; radial basis function networks; three-term control; tracking; velocity control; PID control parameters; genetic algorithm; gradient descent method; injection molding control; nonlinear injection system; ram velocity tracking control; system parameter fluctuation; training RBF neural network; Clustering algorithms; Control systems; Feedforward neural networks; Genetic algorithms; Injection molding; Knowledge acquisition; Mechanical engineering; Neural networks; Three-term control; Velocity control; Genetic algorithm; Gradient descent; Injection system; PID controller; RBF neural network;

fLanguage

English

Publisher

ieee

Conference_Titel

Knowledge Acquisition and Modeling, 2009. KAM '09. Second International Symposium on

Conference_Location

Wuhan

Print_ISBN

978-0-7695-3888-4

Type

conf

DOI

10.1109/KAM.2009.214

Filename

5362315