基于小波包模糊推理的上肢康复机器人智能专家系统

A robotic intelligent expert system for upper-limb rehabilitation based on wavelet packet and Fuzzy reasoning

针对目前上肢运动训练康复机器人缺少对患肢训练模式智能处方诊断的现状，提出了全新的基于小波包模糊推理的上肢康复机器人智能专家系统诊断方法。该方法在智能专家诊断过程中首先根据患肢在水平和垂直方向上的被动运动位置跟踪误差，通过小波包分解提取患肢运动性能特征，然后根据两个方向上的运动性能特征值并运用专家知识，通过模糊逻辑推理诊断该患肢适合的康复运动训练类型。临床实验结果表明，该方法能够较准确地实现不同病情患肢的运动训练模式处方诊断，有助于提高康复机器人的临床智能化水平。

Considering the fact that the existing robot-aided upper-limb rehabilitation systems are incapable of automatically recommending an appropriate training mode for an impaired limb, a novel intelligent expert system diagnosis method is proposed based on wavelet packet and fuzzy logic. When the method is applied to diagnose, the impaired limb＇s movement characteristics are firstly extracted using wavelet packet decomposition according to the passive movement tracking errors in horizontal and vertical directions, and then fuzzy logic involving expert knowledge is used to reason out the appropriate recovery training mode for the impaired limb based on the two-direction movement features. The results of clinical experiment indicate that the proposed method is capable of accurately recommending an appropriate training mode to the impaired limbs with different physical conditions, improving the clinical intelligence for a the robot-aided rehabilitation system.