Control of an arm assistive robot

Abstract

Stroke is one of the leading causes of impairment of one’s upper-limb and deprivation of the ability in conducting activities of daily lives (ADL). Researches have shown that robotic devices can extenuate the limitations of conventional rehabilitation therapy such as labor cost and repetitive motion. Besides, a controller that can provide high transparency, minimum assistance necessary, and adaptive to the user’s capability is crucial for rehabilitation robotic devices. This project aims to develop and implement such a controller for an upper limb assistive robot to achieve smooth and safe human-robot interaction while permitting maximum human engagement. The control strategy developed is an integration of dynamic compensation, assist-as-needed control, and adaptive control. In detail, the dynamic compensation strategy is first used to make the robot transparent to the user. Then, an impedance control method using a virtual tunnel and user-defined tolerance is developed to allow the robot to achieve assist-as-needed function. Lastly, a performance evaluation method is proposed to evaluate user performance, and the robot can adaptively adjust the assistive level based on the user’s performances. Experiments on a healthy subject are conducted to verify the effectiveness of the controller by using a resistive band to limit the subject’s ability to simulate the patient’s condition. The results of these experiments proved the robot can provide an effective and safe user experience.