Finite-time control of uncertain Euler-Lagrange systems using ELM-based velocity observer

This paper is concerned with the finite-time trajectory tracking control of a class of Euler-Lagrange (EL) systems under unknown velocity information, dynamic uncertainties and disturbances. An extreme learning machine (ELM) algorithm is employed to approximate the uncertainties, while an adaptive algorithm is proposed to tune output weights of the ELM, as well as to eliminate the negative effects of residual errors and disturbances. Then an adaptive finite-time ELM-based velocity observer is developed to estimate the unavailable velocity states. Further, based on the estimations and the approximations of model uncertainties, an adaptive finite-time observer-based nonsingular terminal sliding mode (TSM) control strategy is constructed to guarantee the finite-time bounded tracking of the uncertain EL system by using the Lyapunov stability theorem. Simulation results on a robotic manipulator platform demonstrate the efficiency of the developed finite-time observation and control methods.