中文摘要：

针对机器人辅助被动康复训练运动很少考虑训练过程中患肢状态变化的情况，提出了基于患肢状态观察器的被动康复训练运动控制方法。该方法用患肢状态观察器根据位置和速度跟踪误差提取患肢运动特征，运用模糊推理逻辑实时评估患肢物理状态；然后动态规划决策机制根据患肢状态自适应地调整规划运动速度；最后利用位置阻抗控制实现训练运动。定量和定性两类实验结果表明，该方法较传统方法具有更好的跟踪性能和运动平稳性，且能够实现人性化和智能化的康复训练运动。

英文摘要：

Considering that robot-assisted passive rehabilitation-training exercises rarely manage impaired-limb state chan- ges in the training process, a motion control method based on an impaired-limb state observer is proposed for pas- sive rehabilitation training. Firstly, the method uses an impaired-limb state observer to acquire the limb＇ s move- ment characteristics according to position-and-velocity tracking errors, and fuzzy reasoning is used to evaluate the impaired-limb physical state in real time. Then the dynamic programming decision-making mechanism adaptively adjusts the planning movement speed according to the evaluated state. Finally, the training exercise is carried out with position-based impedance control. The results of quantitative and qualitative experiments indicate that the pro- posed method presents a better tracking performance and motion stability than traditional methods, meanwhile, sup- plying more humane and intelligent rehabilitation training.