Barrier Lyapunov Function-Based Fuzzy Adaptive Admittance Control of an Upper Limb Exoskeleton Using RBFNN Compensation

The application of an exoskeleton can reduce the physical effort of wearer in many repetitive industrial tasks. However, the variation of human-robot-environment interaction properties and the uncertainties of dynamic modeling increase the difficulty in accurate coordination control. To overcome this problem, in this article, a fuzzy adaptive admittance control approach is developed in which the exoskeleton coordination control is combined with a barrier Lyapunov function (BLF) and a neural network compensator. The motion signals of interaction force and end-effector velocity are collected and fused into a high-level controller with a fuzzy adaptive admittance model, which is used for the adjustment of cooperation characteristics during work. A BLF-based low-level controller is proposed to improve the position control accuracy, and the system uncertainties are compensated via a radial basis function neural network approximator. The validity of the proposed method is verified via three experiments, including trajectory tracking experiment, cutting operation experiment, and grinding operation experiment. The results showed that the developed control strategy can guarantee the control performance and reduce human muscular power consumption in industrial interaction tasks with different compliance levels.