MODELING AND CONTROL OF NON-LINEAR SYSTEMSUSING NEURO-FUZZY APPROACH

Abstract

In this work a neuro-fuzzy approach is used to model any non-linear data. Fuzzy curve approach is used to know prerequisite parameters to model the system. Back-propagation algorithm is used to properly train the network. The appropriateness of the model is tested with a non-linear data and the model results are compared with actual data. Neuro-fuzzy controller is designed for LOS stabilization for a two axis gimbal system. Implementation in azimuth axis is presented. A conventional compensator designed in [2] is used as training data for neuro-fuzzy controller. Fuzzy logic based controller is implemented on the system. Neuro-fuzzy model algorithm is used in modelling the controller. Step response of the system using the three controllers is implemented in MATLAB and the results are compared.