Design of a bionic-inspired exoskeleton robot for lower limb assist

Abstract

The design of an intelligent exoskeleton robot with pneumatic artificial muscles for human lower limb motion assist using electromyography (EMG) is presented. There are four topics addressed in this paper. Decoding electromyography is the first topic. When muscles are active, they produce an electrical activity. EMG is a record of this electrical activity that reflects human’s movement. Through regression analysis a model is obtained to extract motion commands from EMG. It would be an advantage to employ EMG as a control signal for the exoskeleton control. Second, the pneumatic artificial muscle, air muscle for short, is a simple and powerful actuator. When actuated with compressed air, it contracts and provides a pulling force. As a result of its behavior in a similar way to a biological muscle, air muscle is adapted for a bionic actuator of the assist robot. The force models of air muscles are investigated by experiments in a workbench. Third, for the control of a bionic-inspired robot, the multimodal sensory feedback including EMG and inertial sensors is necessary. By using EMG as a force-proportional measurement between human and robot, a control system combined a sensor-fusion approach and a compliant mechanism enables exoskeleton to carry out human-robot collaboration. Finally, a prototype of power-assist exoskeleton robot for lower limb is completed and evaluated by experiments successfully.