Optimization-based configuration for parallel cable-driven exoskeletons enhancing compatibility

Abstract

Cable-driven exoskeletons have become a vital alternative for poststroke rehabilitation. However, since cables generate only force vectors, the human-machine interaction wrench in parallel cable-driven exoskeletons usually contains the incompatible component that does not contribute to human joint actuation, leading to limited compatibility. This paper proposes a novel optimization-based configuration method to reduce the incompatible interaction wrench in parallel cable-driven exoskeletons. The human-machine interaction wrench in the parallel cable-driven exoskeleton is first modeled, distinguishing between compatible and incompatible components based on the motion characteristics of the human joint. A passive exoskeletal mechanism is introduced to classify the parallel cables into two segments. The wrenches generated by these cable segments are associated exclusively with the compatible and incompatible interaction wrenches, respectively, thereby decoupling these components. Under this configuration, an optimization is performed on unit cable wrenches to reduce the magnitude of the incompatible interaction wrench while preserving the compatible one. Subsequently, the proposed method is specifically applied to a shoulder rehabilitation exoskeleton. The results of simulations and prototype experiments indicate a significant reduction in the magnitude of the incompatible interaction wrench, thus validating the effectiveness of this method in improving human-machine compatibility.