Implementation of Hybrid Neuro-fuzzy controller based on Radial Basis Function Network with gradient descent algorithm for non-linear system
Ravi Kumar Soni, M. J. Nigam, R. Mitra · 2015
The main focus of this paper is on stabilization of non-linear system using proposed Hybrid Neuro-fuzzy Logic Controller based on Radial Basis Function Network (HNFRBFN). A control unit matrix has been introduced according to the non-linearity and uncertainty of the plant. The parameters used to obtain the individual control action in this matrix are optimized by Gradient Descent Algorithm. IF part of the fuzzy rules is approximated using Radial Basis Function (RBF) Network & THEN part of fuzzy rules are synthesized by the control action matrix which incorporates Error-correction algorithm. This error correction algorithm forms the basis of HNFRBFN. The simulation results obtained show that the suggested controller gives promising improvement and better performance.