Model Reference Adaptive Impedance Controller Design For Modular Exoskeleton

This work presents an implementation of a Model Reference Adaptive Controller (MRAC) intended for human and robot cooperation. In order to achieve this cooperation, the plant is composed by the human and exoskeleton together, and the adaptive controller is used to estimate the plant parameters. Simulations showed that the method is effective when dealing with model uncertainties by estimating the plant parameters. Moreover, it showed robustness to different input functions and plant non-linearities. The simulations indicate that this type of controller is suitable to assist human movement by reducing the strength needed for the user to make a movement. In addition, it can estimate the human arm dynamics that is a useful feature in a rehabilitation robot. Experiments with a modular exoskeleton are being carried out in order to validate the simulations.