This article addresses the reference tracking issue of multi-degree-of-freedom robotic exoskeletons under exogenous perturbations. First, by considering the concept of iteration-dependent full-format dynamic linearization, the exoskeleton robot’s dynamics is reformulated as a linear data model in an iterative manner. Then, based upon an iterative sliding variable, a novel data-based model-free adaptive iterative learning integral terminal sliding mode control is designed. The superiority of the proposed study is that the reference tracking problem is just solved by using the measured input/output information of exoskeletons. In addition, its finite-iteration convergence is also affirmed by mathematical analysis. The simulation investigation together with the compared results also clarifies the efficiency of the developed learning-based control algorithm.
This paper presents an adaptive robust control algorithm for the nonlinear dynamics of robot manipulators with unknown backlash in gears. The basic nonlinear model of a serial manipulator robot is used for the controller design, and this is combined with the nonlinear proposed dead zone model, based on the input and output torque. The main idea of providing this model is to achieve a dynamic model of the system considering the backlash of the robot joint gears, and having less complexity such that the developed controller does not need the inverse backlash model. The adaptive robust controller is developed, without using the inverse backlash model. The proposed controller is designed based on an unknown dead zone parameter and it guarantees the stability and path tracking of the robot trajectory with unknown dead zone parameter in the desired range. Numerical simulations are conducted to show the effectiveness of the proposed controller. Finally, the efficiency and capability of the proposed controller in dealing with the unknown backlash nonlinearities in gears of the manipulator are demonstrated by experimental results with a five-bar manipulator.