外骨骼
自适应控制
上肢
物理医学与康复
计算机科学
医学
康复
上半身
弹道
生物力学
控制理论(社会学)
康复机器人
控制(管理)
假肢
下肢
上下界
工程类
自适应滤波器
矫形学
作者
Jiazhen Xu,Haoping Wang,Yang Tian,Yangchun Wei
标识
DOI:10.1109/thms.2026.3664850
摘要
Assist-as-needed (AAN) control for upper-limb rehabilitation exoskeletons is crucial for promoting active patient participation and facilitating neuroplasticity after a stroke. However, a significant challenge lies in designing controllers that can both seamlessly switch assistance levels in response to a patient's fluctuating ability and rigorously enforce safety-critical position constraints across different interaction modes. This study proposes a Multi-Mode Assist-as-Needed (MMAAN) control scheme to address these issues. The proposed framework features an outer loop that uses a virtual tunnel and a novel, continuously differentiable Robot-Assisted Level Metric (RALM) to ensure smooth transitions between Human-Dominated, Human-Robot shared, and Robot-Dominated modes. The inner loop implements a Performance-Varying Barrier Lyapunov Function controller, whose constraint boundaries dynamically adapt based on the RALM to guarantee safety and tracking accuracy. Experimental results with human participants demonstrate that the MMAAN controller automatically adjusts assistive force according to the user's performance while Maintaining tracking errors within prescribed boundaries. This approach offers a robust framework for personalized and safe robotic rehabilitation, potentially enhancing clinical outcomes by adapting to the unique needs of each patient.
科研通智能强力驱动
Strongly Powered by AbleSci AI