Authors: Ashwani Kharola; Pravin Patil
Addresses: Department of Mechanical Engineering, Graphic Era University, Dehradun-248002, India ' Department of Mechanical Engineering, Graphic Era University, Dehradun-248002, India
Abstract: This paper presents a MATLAB/Simulink-based offline control of highly nonlinear, multivariable and complex ball and beam system. The objective is to control both beam orientation and ball position using PID-based ANFIS controllers. The simulation results of PID controller were used for training of ANFIS controller. The FIS for ANFIS controller were designed using three gbell membership functions. The simulation results of PID-based ANFIS controller showed excellent learning ability and control of ball and beam system as compared to conventional PID controller.
Keywords: ball and beam systems; PID control; adaptive neuro-fuzzy inference system; ANFIS; fuzzy logic; FIS; gbell membership function; MATLAB; Simulink; surface viewer; training error; epochs; controller design; neural networks; offline control; nonlinear systems; beam orientation; ball position; simulation.
International Journal of Advanced Mechatronic Systems, 2016 Vol.7 No.1, pp.24 - 34
Received: 04 Apr 2016
Accepted: 06 Jul 2016
Published online: 04 Oct 2016 *