Publication: Pembangunan kawalan suai dalam-talian untuk sistem kawalan kelajuan motor dalam matlab melalui basic stamp
Loading...
Date
2005-03-01
Authors
Rifin, Rozi
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This project presents a Local Output Local Recurrent Globally Feedforward Neural Network (LOLRGF) for real time control implementation which is DC motor
speed control. The on-line indirect adaptive control scheme is developed and neural network is used as compensator in the closed loop system to control nonlinear processes. The LOLRGF neural network was trained using Recursive Prediction Error (RPE) algorithm in order to output process follow the behavior of set point tracking beside to minimize an offset as small as possible. A continuously stirred tank reactor (CSTR) is used to develop, modeling and designing neural network compensator in simulation to determine the best output process according to network size, learning parameters and network connection. The DC motor speed control system will be adapted in the model to substitute CSTR process and observation regarding to output system behavior are implemented for real time application. In this investigation, MATLAB/SIMULINK software was used in control schemes development, modeling, analysis and design. This software also acts as controlling medium for DC motor speed control system which use Basic Stamp 2 as microcontroller.