Fuzzy Inference-Based Driving Control System for Pushrim-Activated Power-Assisted Wheelchairs Considering User Characteristics

Pushrim-activated power-assisted wheelchairs that assist the driving power by electric motors are effective in the daily lives of physically disabled people in terms of expanding the range of daily activities and comfortable driving even on rough and uphill roads. Meanwhile, users with unbalanced pushing power of the right and left arms due to some disability have difficulty controlling the wheelchair as he or she requests. For example, users with a lower-right arm strength have difficulty simultaneously driving forward and to the left. Therefore, a power-assist control system that considers the wheelchair operation characteristics of the individual user should be designed. The present article proposes a driving control system of pushrim-activated power-assisted wheelchairs with the target position estimation based on a learning model of individual operation characteristics. An individual estimation model based on fuzzy inference is constructed based on the right- and left-pushing torque information, which is properly revised to improve the position estimation accuracy. In addition, an operation training system with visual presentations to improve the proficiency of the user is effectively used. Experimental results using trial subjects, including a physically disabled person with arthrogryposis multiplex congenita, show that assisted driving intended by users can be realized, which verifies the effectiveness of the proposed driving control system and training system.