Sensory controlled intelligent assistant system for walking rehabilitation

This paper presents a sensory controlled intelligent assistant system for walking rehabilitation. Several pressure sensors are used to measure pressure distribution by the foot during walking, and an encoder is used to record trainee angle of knee joint. The gait phases are detected by the data that come from pressure and angular sensors. To realize the system in practical applications, we represent the rehabilitation expert knowledge by fuzzy membership functions and rules. If the trainees gait phase does not represent as usual phase, we will give him an extra force to correct his phase cycle through the fuzzy inference engine. 30 days as a course of treatment is set, trainee walking phases are traced and recorded into database during this period. In order to reduce the quantity of therapist's workload, this system is designed to recognize every trainee automatically who is in the database. The recognition method is based on neural network that inputs the angle of joint and foot pressure distribution directly. This system is able to adapt to the trainee walking phase to increase the training effect and recognize trainee automatically.