Adaptive Terminal Sliding Mode Tracking Control for Rigid Robotic Manipulators with Uncertain Dynamics.

Abstract

A robust adaptive terminal sliding mode tracking control is proposed for rigid robotic manipulators. First, it is shown that, when a rigid robotic manipulator is treated as a partially known system, the system uncertainty is not only related to the model properties but also to the structure of the controllers. It is also proved that, if the control input vector does not contain the acceleration signals, the system uncertainty is upper bounded by a positive function of the position and velocity measurements. The property is very useful for the design of robust and adaptive controllers where only position and velocity measurements are available. Second, an MIMO terminal sliding mode is defined for the error dynamics of the robot control system, and an adaptive mechanism is introduced to estimate the upper bounds of system uncertainty in the Lyapunov sense. The estimates are then used as controller parameters so that the effects of uncertain dynamics can be eliminated and a finite time error convergence can be guaranteed. A simulation is given in support of the proposed control scheme.