Direct Neuro – Adaptive Controller Based on Finite Recurrent Back Propagation (FRBP) Network

Abstract

For nonlinear systems, the method of directly adjusting the parameter of the controller using the adaptive control are not available. This is because the unknown nonlinear system lies between the controller and the output error. This paper explains the abilities of FRBP network to capture the nonlinearity and this will often the ability to overcome the problem of direct adaptive control in nonlinear system. Two neural networks are used as a controller for a pitch channel for BTT missile control system, the simulation results shown the ability of FRBP for capture the nonlinearity rather than fast BP.