A Cable-Driven Three-DOF Wrist Rehabilitation Exoskeleton With Improved Performance

• Published in: Frontiers in neurorobotics Vol. 15; p. 664062
• Main Authors: Shi, Ke, Song, Aiguo, Li, Ye, Li, Huijun, Chen, Dapeng, Zhu, Lifeng
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Research Foundation 08.04.2021; Frontiers Media S.A
• Subjects: Algorithms; Bandwidths; cable-driven robot; Cables; Control algorithms; distributed drive system; Efficiency; Exoskeleton; human-robot interaction; mechanism design; Neuroscience; Rehabilitation; rehabilitation robot; Robots; Stroke; Wrist; distributed drive system; human-robot interaction; mechanism design; cable-driven robot; rehabilitation robot
• ISSN: 1662-5218; 1662-5218
• DOI: 10.3389/fnbot.2021.664062

This paper developed a cable-driven three-degree-of-freedom (DOF) wrist rehabilitation exoskeleton actuated by the distributed active semi-active (DASA) system. Compared with the conventional cable-driven robots, the workspace of this robot is increased greatly by adding the rotating compensation mechanism and by optimizing the distribution of the cable attachment points. In the meanwhile, the efficiency of the cable tension is improved, and the parasitic force (the force acting on the joint along the limb) is reduced. Besides, in order to reduce the effects of compliant elements (e.g., cables or Bowden cables) between the actuators and output, and to improve the force bandwidth, we designed the DASA system composed of one geared DC motor and four magnetorheological (MR) clutches, which has low output inertia. A fast unbinding strategy is presented to ensure safety in abnormal conditions. A passive training algorithm and an assist-as-needed (AAN) algorithm were implemented to control the exoskeleton. Several experiments were conducted on both healthy and impaired subjects to test the performance and effectiveness of the proposed system for rehabilitation. The results show that the system can meet the needs of rehabilitation training for workspace and force-feedback, and provide efficient active and passive training.