Model-free Adaptive Trajectory Tracking Control of Robotic Manipulators with Practical Prescribed-time Performance

Abstract

Considering accurate model information of robotic manipulators is hard to obtain in real applications, this article proposes a model-free trajectory tracking control scheme for the robotic system with input saturation, where a pre-defined tracking precision within prescribed time under any initial conditions is ensured. Firstly, to avoid a sharp corner when control input exceeds constraint, a new smooth function is used to approximate the saturated control torque and then a new robotic system model is built. Based on the rebuilt system, a time-delay estimation (TDE) method is used to estimate the lumped uncertainty containing unknown dynamics and external disturbance. Furthermore, a model-free adaptive controller is developed to ensure the tracking error converge to a preset small residual within given time, where adaptive law and auxiliary system are used to cope with the TED estimation error and control input saturation, respectively. With the developed practical prescribed-time function, there is no constraint on the initial value of the tracking error, thus global stability is guaranteed. Finally, a planar two-link robotic manipulator is simulated to show the effectiveness of the developed control scheme.