Neuro-adaptive dynamic integral sliding mode control design with output differentiation observer for uncertain higher order MIMO nonlinear systems

• Main Authors: Khan, Qudrat, Akmeliawati, Rini
• Format: Article
• Language: English; English; English
• Published: Elsevier 2017
• Subjects: TJ212 Control engineering
• Online Access: http://irep.iium.edu.my/53663/1/53663_Neuro-adaptive%20dynamic%20integral%20sliding%20mode.pdf; http://irep.iium.edu.my/53663/2/53663_Neuro-adaptive%20dynamic%20integral%20sliding%20mode_SCOPUS.pdf; http://irep.iium.edu.my/53663/3/53663_Neuro-adaptive%20dynamic%20integral%20sliding%20mode_WOS.pdf; http://irep.iium.edu.my/53663/; http://www.sciencedirect.com/science/article/pii/S0925231216314382

This paper proposes a practical design method for the robust control of a class of MIMO nonlinear plants operating under model uncertainties and matched disturbances where the only available information for feedback are the outputs of the plant. A neural networks based dynamic integral sliding mode control (NNDISMC) with output differentiator observer is developed for the considered class. This NNDISMC approach utilizes the robust output differentiation observer for the higher derivative estimation and neural networks to estimate the nonlinear functions which are assumed unknown. Having estimated the unknown derivatives and uncertain functions, an integral manifold based on the estimated states is designed and a control law is proposed which confirms the sliding mode enforcement across the designed integral manifold from the very start of the process. The overall robustness of the controller is guaranteed by using the neural networks, differentiator observer and dynamic integral control law in a closed loop. The closed loop stability analysis is presented in detail, and the asymptotic convergence of the system states to the equilibrium is confirmed. The proposed method is very practical and plays a very significant role in the robust control of electromechanical systems, such as robotic manipulators, unmanned air vehicles and underwater vehicles. The simulation results on a robotic manipulator are presented to demonstrate the effectiveness of the proposed method.