A task-oriented robotic hand rehabilitation system for post-stroke recovery
Toh, Chen Koon
Date of Issue2015
School of Mechanical and Aerospace Engineering
In emulating physical therapists’ role in occupational therapy, the methodology of exoskeleton based rehabilitation robot is widely researched in facilitating the rehabilitation and motion re-learning of post stroke patients. The research focuses on the design and simulation of a hand rehabilitation exoskeleton robot. Unlike the exoskeleton robots for arm and shoulder rehabilitation, hand exoskeleton systems are required to handle the multiple joints on hand with four times the degree of freedom in a confined space. In addition, exoskeleton robot as a physical wearable system requires careful design of human machine interface, in order to ensure the patient's safety and to avoid discomfort or injuries on his hand. After reviewing different rehabilitation techniques used in post-stroke recovery, the rehabilitation exoskeleton is designed to provide task oriented training guided with Bobath’s rehabilitation technique. The robot is designed to actuate and drive the patient’s hand in natural motion, while preventing the patient from developing wrong muscle tone. In order to better prepare the patient to regain his self-care abilities, the rehabilitation training with the robot is based on activities of daily living (ADL) exercise, which revolves around manipulation of daily objects. With reference to the anatomy and kinematic model of human hand, a hand rehabilitation robot utilizing an innovative arc-gear mechanism is designed and built. As the exoskeleton has to be designed at the dorsal side of hand due to space constraint, it is essential for the exoskeleton structure to provide a guided movement which is concentric to the human joints. The arc-gear mechanisms driven by DC motors aimed to replicate an accurate and natural hand movement to exercise patients’ hand. State of the art from other institutions will also be discussed to justify the strength of the designed machine. The workspace of the human hand is plotted in the 3D CAD software to verify the working range of the exoskeleton. Amplifiers for the DC motors have been designed and built by utilizing the control signal from NI motion control hardware. The ADL training chosen for the research is reach-to-grasp. The user is required to reach for a mug and grasp it, before retrieving it back to the original position. The trajectories for the reaching and grasping movements are carefully designed to iii provide the natural motion profile. Position control is introduced at task space and joint space level to monitor the motion of exoskeleton. As conclusion to the research, the model is simulated with the NI Softmotion, whereby the motion command is generated from the Labview software to move the 3D model of the hand exoskeleton in Solidworks software. The results from the motion are then compared to the actual reach-to-grasp trajectories performed by healthy persons as reported in various literatures. The motion profile of the planned trajectories actuated by the designed exoskeleton is verified to be similar to the natural reach-to-grasp trajectories.