垂足是中風患者常見的動作障礙，主要由背屈肌群無力造成，本研究的目的為利用EMG訊號回授控制使患者踝關節的主動扭矩可在背屈運動過程中維持定值，訓練中風患者控制其背屈肌群肌力。

在復健機器人系統中取得主動扭矩的方法有兩種：一為分析機器人與受測者踝關節之間的作用力，由扭力計測得之扭矩分解出主動扭矩，此外，亦可利用脛骨前肌的EMG訊號估測主動扭矩。兩種方法所取得的主動扭矩並不相同，由扭力計所估測的主動扭矩為所有肌肉作用在踝關節上的合力矩，而EMG訊號估測的為背屈肌群所產生的主動扭矩。本研究發展一新的主動扭矩控制方法，可透過誤差權重調變上述兩種主動扭矩在回授控制中的比重，以適應病人之肌肉控制能力。由於肌電訊號的變動較大，訊號的不平順會使得中風病人難以適應由機器人所提供的力量，因此我們將由肌電訊號估測之主動扭矩通過適應濾波器，在許可範圍內調整估測器的頻寬，使估測之主動扭矩較為平順，增加系統之可適應性。

本研究以兩位常人及五位中風病人進行實驗，發現在以扭力計估測之主動扭矩為回授時，對受測者的運動干擾較小，而隨著誤差權重的增加，外加扭矩對受測者的影響也變大。結果顯示不論是在扭力計回授模式或者在EMG回授模式下，受測者的主動扭矩在背屈運動過程中可維持定值。

Foot drop is the most common movement disorder of the strokes, which is mainly due to the weakness of dorsiflexors. The goal of this study is to develop a method of controlling the active ankle torque to maintain constant by using EMG feedback to train the dorsiflexors of the strokes.

Two approaches to obtain the active torque of ankle were adopted. One is to analyze the reactive torque between the robot and subject’s ankle and obtains the relationship between active torque and the measured reactive torque. The other is to estimate the ankle active torque from the EMG signals of tibialis anterior. These two kinds of active torques are different in that the active torque detected by the torque sensor is the resultant torque from dorsiflexors and plantarflexors. While the active torque estimated from the EMG is sole from the dorsiflexors. In this work the weighted error of above active torques was employed to develop a new active torque control system to adapt to various control capability of the patients. Since the variations of EMG-estimated torque are large, the robot may generate a non-smooth assisting torque which may interfere in the subjects’ tracking control. An adaptive filter was employed to tune the bandwidth of the torque estimator to generate a smooth command.

Two normal subjects and five stroke subjects were recruited in this study. The results revealed that the feedback based on the active torque estimated from the reactive torque has less interference to the active tracking in both groups of subjects. Increasing the weight on error from EMG-estimated active torque would induce higher interference. The active torque of all subjects can be controlled at the desired level during the robot-assisted tracking with both torque and EMG feedback.

1. Yeh C.-Y., Chen J.-J.J. and Tsai K.-H., Quantifying the effectiveness of the sustained muscle stretching treatments in stroke patients with ankle hypertonia. J. Electromyogr. Kinesiol., 2007. 17(4): pp. 453-461.
2. 陳佳萬 朱耀堂.郭有順. 腳踝關節復健機.中華民國專利,專利公告號: 555555. 2003.
3. Bressel E. and McNair P.J., The effect of prolonged static and cyclic stretching on ankle joint stiffness, torque relaxation, and gait in people with stroke. Phys. Ther., 2002. 82(9): pp. 880-887.
4. Zhang L.-Q., Chung S.-G., Bai Z.-Q., Xu D.-L., van Rey E.M.T., Rogers M.W., Johnson M.E. and Roth E.J., Intelligent stretching of ankle joints with contracture/spasticity. IEEE Trans. Neural Syst. Rehabil. Eng., 2002. 10(3): pp. 149-157.
5. Zhang L.-Q., Selles R.-W., Li X.-Y., Lin F., Chung S.-G. and Roth E.J. Feedback-controlled and programmed stretching of the ankle plantarflexors and dorsiflexors in stroke: Effects of a 4-week intervention program. 50th Annual Meeting of the Orthopaedic-Research-Society. San Francisco, CA, 2004: pp. 2330-2336.
6. Wheeler W. J., Krebs H. I., Hogan N., An ankle robot for a modular gait rehabilitation system. Proceedings of 2004 IEEE/ASJ International Conference on Intelligent Robots and Systems, 2004: pp. 1680-1684.
7. Ju M.-S., Lin C.-C.K., Lin D.-H., Hwang I.-S. and Chen S.-M., A rehabilitation robot with force-position hybrid fuzzy controller: Hybrid fuzzy control of rehabilitation robot. IEEE Trans. Neural Syst. Rehabil. Eng., 2005. 13(3): pp. 349-358.
8. Kung P.-C., Ju M.-S. and Lin C.-C.K. Design of a forearm rehabilitation robot. IEEE 10th International Conference on Rehabilitation Robotics. Piscataway, NJ, United States, 2007: pp. 228-233.
9. Lin C.-C.K., Ju M.-S., Chen S.-M. and Pan B.-W., A specialized robot for ankle rehabilitation and evaluation. Journal of Medical and Biological Engineering, 2008. 28(2): pp. 79-86.
10. Cheng H.-S., Ju M.-S. and Lin C.-C.K., Improving elbow torque output of stroke patients with assistive torque controlled by emg signals. J. Biomech. Eng.-Trans. ASME, 2003. 125(6): pp. 881-886.
11. Stein J., Narendran K., McBean J., Krebs K., and Hughes R., Electromyography-controlled exoskeletal upper-limb-powered orthosis for exercise training after stroke Am. J. Phys. Med. Rehabil., 2007. 86: pp. 255-261.
12. Song R., Tong K.-Y., Hu X.-L. and Li L., Assistive control system using continuous myoelectric signal in robot-aided arm training for patients after stroke. IEEE Trans. Neural Syst. Rehabil. Eng., 2008. 16(4): pp. 371-379.
13. 潘柏瑋. 中風病患踝關節復健用機器人之研究. 國立成功大學 機械工程學系 碩士論文 2006.
14. Netter F.H., Interactive atlas of human anatomy, Ciba Medical Education & Publications, 1995.
15. Winter D.A., Biomechanics and motor control of human movment. 2nd ed. 1990: John Wiley & Sons, Inc.
16. Hogan N. and Mann R.W., Myoelectric signal processing: Optimal estimation applied to electromyography - part 1: Derivation of the optimal myoprocessor IEEE Trans. Biomed. Eng., 1980. 27(7): pp. 382-395.
17. Park E. and Meek S.G., Adaptive filtering of electromyographic signal for prosthetic control and force estimation. IEEE Trans. Biomed. Eng., 1995. 42(10): pp. 1048-1052.