Fuzzy Control Method for an Intelligent Walking Training Robot for User Fall Prevention

In this paper, a fuzzy control method is proposed for user fall prevention during walking training of the intelligent walking training robot (IWTR). In clinical rehabilitation sites, an experienced human caregiver can always assist patients with walk training and prevent them from falling risks by providing observational and force interaction information. However, for a robot to detect the user's falling risk and make the appropriate movement to help the user to obtain balance is challenging. In this study, we investigated a fuzzy control method for motion control of an IWTR to implement a falling prevention strategy that would be the same as provided by human experts during walking rehabilitation. By extracting the knowledge of human experts on fall prevention as fuzzy rules, the IWTR can determine the appropriate falling prevention motion according the information from a monocular camera and four force sensors. The monocular camera is applied to sense the relative position of the user to the IWTR. The force sensors are placed under the armrest of the IWTR to detect the interaction force between the user and IWTR. The distance and interaction force information are taken as the input for the fuzzy controller, and the motion velocity can be determine to control the IWTR to prevent users from falling down similar to human rehabilitation experts. Finally, experiments were implemented for verification. The result showed the effectiveness of the proposed method for fall prevention.