The Design of Stroke Rehabilitation Using Artificial Intelligence K.A.K.I (Kinesthetic Augmented Kinematic Inference)

Abstract

Stroke is a major cause of disability worldwide that affects many people every year. Stroke rehabilitation is a process that helps stroke patients regain their lost function and improve their quality of life. However, the recovery process varies widely depending on the severity of stroke and other factors such as age, health and type of stroke. Many elderly patients face difficulties in attending rehabilitation centers due to various factors such as cost, distance and congestion. Therefore, this paper proposes methods to help stroke patients do rehabilitation exercises at home using the latest technology. Our project consists of interactive exercises that are customized to the skill level of the patients, hardware sensor inputs that can measure the strength of the hand movement of the patients, embedded processing board with camera that can detect and guide the movement of the patients and machine learning using convolutional neural network (CNN) that can analyze the movement data and provide feedback and motivation to the patients. The effectiveness of the proposed system is evaluated by the improvements in patients’ conditions through pre- and post-exercise tests. Overall, our kinesthetic augmented kinematic inferencing methods appear to be more effective than conventional methods for post-stroke rehabilitation. This project demonstrates a promising solution to enhance stroke rehabilitation, recovery and quality of life.