Adaptive composite neural network disturbance observer-based dynamic surface control for electrically driven robotic man
- PDF / 3,296,584 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 86 Downloads / 196 Views
(0123456789().,-volV)(0123456789(). ,- volV)
ORIGINAL ARTICLE
Adaptive composite neural network disturbance observer-based dynamic surface control for electrically driven robotic manipulators Jinzhu Peng1,2
•
Shuai Ding1 • Rickey Dubay2
Received: 27 February 2020 / Accepted: 24 September 2020 Ó Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract This paper presents an adaptive backstepping control scheme for electrically driven robotic manipulator (EDRM) system with uncertainties and external disturbances by using neural network disturbance observer (NNDO) and dynamic surface control (DSC) design technique. NNDO is employed to estimate the uncertainties and external disturbances such that the priori information of the unknown dynamics will not be needed. To overcome the problem of ‘‘explosion of complexity’’ inherent in the backstepping design method, the DSC technique is integrated into the adaptive backstepping control design framework, where the NNDOs with adaptive composite law are utilized to compensate the uncertainties and external disturbances of EDRM. Based on the Lyapunov stability theory, it can be proven that the closed-loop system is stable in the sense that all the variables are guaranteed to be uniformly ultimately bounded. The results of simulation and experimental tests demonstrate the approximation capability of NNDO and the effectiveness of the proposed adaptive DSC scheme. Keywords Neural network Disturbance observer Dynamic surface control Electrically driven Robotic manipulator Backstepping control
1 Introduction The high-precision control of robotic systems is always aspired by researchers; lots of powerful control methods have been therefore applied, including computed torque control methods [1, 2], fuzzy and/or neural network (NN)based control methods [3–6], adaptive control methods [7, 8], robust control methods [9, 10] and so on. The above investigations are mainly focused on the torque control for the rigid-link robotic manipulators. It should be pointed out that robotic manipulators are generally driven by the motors in most applications, the control inputs for robotic & Jinzhu Peng [email protected] Shuai Ding [email protected] Rickey Dubay [email protected] 1
School of Electrical Engineering, Zhengzhou University, Zhengzhou 450001, Henan, People’s Republic of China
2
Department of Mechanical Engineering, The University of New Brunswick, Fredericton, NB E3B5A3, Canada
manipulators are the voltages of the motors, and the relationship between the joint subsystem and the motor subsystem should not be ignored [11]. However, the robotic control systems will become more complex and difficult provided that the relationship between motor voltages and torques is considered. Therefore, lots of powerful control methods have also been applied to the electrically driven robotic manipulators (EDRM) to achieve the certain performances [11–18]. Fateh et al. presented the robust decentralized control with adaptive fuzzy estimation of uncertainty [11] and the robust di
Data Loading...
