Development of a whole-body walking rehabilitation robot and power assistive method using EMG signals

In this paper, we describe the development of DDGo Advanced, a whole-body walking rehabilitation robot designed as an end-effector type, along with its three walking training methods. Most of the walking rehabilitation robots currently on the market have been developed only to rehabilitate the lower limbs of patients. However, actual human walking performs a whole-body movement with an arm swing. To this point, DDGo Advanced has been developed to provide a natural and correct walking motion of a healthy person by using both a lower-limb rehabilitation part for walking motion and an upper-limb rehabilitation part for the arm swing motion. Using DDGo Advanced, three walking training modes using a parallel hybrid driving unit are proposed. The first is Passive mode in which the robot fully supports the user walking motion. The second is electromyography (EMG)-based Active-assist mode, which checks the muscle signals of the user’s healthy leg and determines the user’s walking intention. It then automatically performs acceleration and deceleration of the driving motor. The third is Active mode with specific angles assist which allows the user to perform walking rehabilitation training with mostly the power of the user and only minimal assistance from the robot. In addition, we adopted a device capable of automatically changing the step length according to the walking speed. Finally, we verified with experiments that the walking motion of the user who performed whole-body walking rehabilitation was more similar to the walking of a healthy person than the walking motion of the user who performed only the lower-limb walking rehabilitation. Moreover, the muscle-assistance performance of each of the three modes was verified by experiments using EMG measurement.