Intention Estimation and Walking Assistance for Severe Paraplegic Patients Having Difficulty in Controlling Neuromuscular Activity

Abstract

Recent studies have revealed the importance of voluntary exercise in rehabilitation training after spinal cord injury and cerebrovascular diseases. However, the voluntary exercise has been an impossible task for severe paraplegic patients who cannot move their lower limbs by themselves. If a robot suit can assist walking of such patients by estimating their intention from neuromuscular activities, it can provide an opportunity for voluntary exercise to them. Because the patients have difficulty in controlling neuromuscular activities, the robot suit is required to estimate the intention from deficient neuromuscular activities and to execute gait motions in place of their paralyzed limb muscles. The purpose of this study is to assist gait of severe paraplegic patients with the robot suit by estimating their intention of swinging legs. To this end, we developed a method to estimate the intention of swinging leg from both deficient bioelectrical signals and foot-ground reaction force. In order to move a patient's paralyzed limbs, we combined the estimation method with an autonomous control method for the robot suit. Walking trials on a treadmill with partial body weight support were performed with two complete paraplegic patients. The robot suit was able to estimate the patients' intention from deficient bioelectrical signals and to move their lower limbs. Consequently, the robot suit succeeded in assisting gait of the patients on the treadmill. This result shows that the robot suit can provide an opportunity for voluntary exercise to severe paraplegic patients having difficulty to move their lower limbs.