Two-stage model free fuzzy adaptive controller for multiplicative disturbance rejection

Article type: Research Article

Authors: Kadri, Muhammad Bilal^{; *}

Affiliations: Department of Mechatronics, College of Engineering, Karachi Institute of Economics and Technology, Karachi, Pakistan

Correspondence: [*] Corresponding author. Muhammad Bilal Kadri, Mechatronics Department, College of Engineering, Karachi Institute of Economics & Technology (PAF-KIET), Karachi, Pakistan. Tel.: +92 213 5091114; E-mail: bilal.kadri@pafkiet.edu.pk.

Abstract: Two-stage model free fuzzy adaptive control (MFFAC) has been proposed which rejects the multiplicative input disturbance while also reducing the unnecessary control activity arising due to the sensor noise. Noisy sensor measurements of the input disturbance severely degrades the control performance and might destabilize the control loop. In this work, a novel approach has been proposed which initially develops a model of the multiplicative input disturbance whereas in the second stage the measurement uncertainty is modeled. The uncertainty is ultimately represented as the fuzziness of the control signal. Adaptive conditional defuzzification is employed which utilizes the fuzziness resulting in reduced control activity. The proposed control scheme has shown good control performance in cases where the plant dynamics varies and the sensor noise in the control loop cannot be simply rejected by low pass filtering. Simulation and experimental results adequately validate the effectiveness of the proposed approach.

Keywords: Model free neuro-fuzzy controller, noise multiplicative disturbance rejection, adaptive conditional defuzzification

DOI: 10.3233/JIFS-171548

Journal: Journal of Intelligent & Fuzzy Systems, vol. 36, no. 3, pp. 2907-2922, 2019