A Study on Wearable Power Assist Robot Controlled Based on Biological Signals

为了辅助老年人、残疾人等身体虚弱者的日常活动或康复，对直接附着在人体上并辅助人体运动的外骨骼（可穿戴）机器人进行了研究。在本研究中，动力辅助是基于反映人体运动意图的生物信号肌电图（electromyogram， EMG）信号来实现的，这类动力辅助外骨骼机器人必须基于用户的运动意图来实现自动动力辅助。虽然利用肌电信号实现外骨骼机器人的动力辅助并不容易，但通过提出模糊神经控制，解决了基于肌电信号的动力辅助的许多问题。模糊神经控制是一种能够处理不确定信号的灵活模糊控制与具有自适应/学习能力的有效神经控制的结合。开发了一种用于辅助人类上肢运动（垂直和水平肩部屈伸运动和肘部屈伸运动）的三维外骨骼机器人。本研究考虑取物运动和进食运动，开发了同时辅助肘关节屈伸运动和前臂旋前/旋后运动的2DOF外骨骼机器人和同时辅助肩膀和肘关节屈伸运动的4DOF外骨骼机器人。此外，考虑到实际应用，提出了一种移动外骨骼机器人，即电动移动椅和动力辅助外骨骼机器人的组合。实验验证了所提出的外骨骼机器人的有效性。

A study on exoskeleton (wearable) robot which is directly attached on human body and assists the human motion is carried out in order to assist daily activity or rehabilitation of physically weak persons such as elderly person or disabled person. In this study, power assist is performed based on electromyogram (EMG) signals which are biological signals reflecting human motion intention, since automatic power assist must be realized based on user's motion intention in this kind of power assist exoskeleton robot. Although it is not very easy to realize power assist with exoskeleton robot using EMG signals, many problems of the EMG-based power assist have been cleared by proposing fuzzy-neuro control which is a combination of flexible fuzzy control that is able to deal with uncertain signals and effective neuro control that has adaptive/learning ability.A 3DOF exoskeleton robot had been developed to assist human upper-limb motion (vertical and horizontal shoulder flexion/extension motion and elbow flexion/extension motion). In this study, a 2DOF exoskeleton robot that assists elbow flexion/extension motion and forearm pronation/supination motion simultaneously and a 4DOF exoskeleton robot that assists 3DOF shoulder motion and elbow flexion/extension motion simultaneously have been developed considering object taking motion and eating motion. Furthermore, a mobile exoskeleton robot, a combination of an electric mobile chair and a power assist exoskeleton robot, has been proposed considering practical application. The effectiveness of the proposed exoskeleton robots has been evaluated by experiment.