Recurrent fuzzy neural network adaptive hybrid control for linearized multivariable systems

Article type: Research Article

Authors: Lin, C.M. | Chen, C.H. | Lee, Y.F.

Affiliations: Department of Electrical Engineering, Yuan-Ze University, 135 Far-East Road, Chung-Li, Tao-Yuan, 32026, Taiwan

Note: [] Corresponding author. Tel.: +88 6 3 4638800 ext 2418; Fax: +88 6 3 4639355; E-mail: [email protected]

Abstract: This paper develops a design method of recurrent fuzzy neural network (RFNN) based adaptive hybrid control for multi-input multi-output (MIMO) linearized dynamic systems. This hybrid control system consists a feedback controller and an adaptive RFNN controller. The feedback controller is used as a basic stabilizing controller to stabilize the nominal system and the adaptive RFNN controller is used to deal with unknown part of system dynamics. The adaptive laws of RFNN are derived based on the Lyapunov function so that the stability of the system can be guaranteed. Finally, the proposed control system is applied to an F-16 flight control system. Simulation results demonstrate that the developed control system can achieve favorable tracking performance even with some failure of the flight control system.

Keywords: Adaptive control, recurrent fuzzy neural network, F-16 aircraft

Journal: Journal of Intelligent & Fuzzy Systems, vol. 17, no. 5, pp. 479-491, 2006