Development of a biofeedback control system for electric stimulation using adaptive filtering
表面肌電訊號 ； 經皮神經電刺激 ； 生物回饋 ； 自適性濾波器 ； SEMG ； adaptive filter ； TENS ； bio-feedback
|Volume or Term/Year and Month of Publication||
|Academic Degree Category||
摘要 現今市面上的多功能電刺激器，以表面電極藉由不同頻率、強度的脈衝波刺激肌肉，達到阻斷、舒緩肌肉的效果。但由於無法自動調整刺激而產生刺激過度或刺激不足的現象，因此若能有效依據肌肉反應來調整刺激的強度，則對此現象將有大幅度的改善。 本研究建構一以PC為平台，以LabVIEW為即時分析、控制介面的肌電生物回饋系統，利用受電刺激時的人體表面肌電訊號（SEMG）為訊號源，經過LMS自適性濾波器與blocking處理去除殘餘的刺激訊號後，計算出肌電訊號的功率值，經過生物回饋的機制進而調控刺激訊號的強度，如此週而復始達到使肌電功率維持在一定值的目的。本研究中共設計三項實驗，其中實驗一、二為確立肌電訊號功率與刺激訊號強度之間的關係；實驗三則利用實驗一、二的結果完成系統的建構與驗證。 經由實驗一、二分析後得知，當刺激訊號強度增加時，肌電訊號功率亦成線性增加的趨勢。利用此結果在實驗三中，完成肌電生物回饋控制系統，並驗證證明此系統確實可依SEMG功率調整刺激強度，並控制SEMG功率在特定的範圍內，使肌肉維持一定程度的收縮狀態。 整體而言，本系統已完成初步的系統架構，但本系統中的肌電訊號功率仍受到許多不確定的因素所影響，包含在刺激頻率較高時所引起的的肌肉反應（muscle response）是否該列入考量。同時，若能改進回饋機制有效的預測下次的肌電功率，則可大大提升整體系統的效能與穩定性。
Abstract There are many multi-functional electric stimulators that deliver pulse wave of different frequencies or intensities through the surface electrodes to stimulate muscle and obtain the effect of blocking and relaxing muscle. However, most of them couldn’t automatically adjust the stimulation so that the stimulation is either too little or too much. This can be improved if the intensity of stimulation can be adjusted according to the muscle response. In this study, a biofeedback system was developed using PC and LabVIEW to achieve real-time analysis. The surface electromyogram (SEMG) during stimulation was processed using LMS adaptive filter and blocking process in order to delete the remnant of stimulation signal. The SEMG power was then calculated to control the intensity of stimulation signal such that the SEMG power can be maintained in the designated level. Three experiences were conducted. In the first and second experiences, the relationship between SEMG power and intensity of stimulation was confirmed. And, in the third experience, the system was built and verified. The results of first and second experiences indicate that there is linear relationship between SEMG power and the intensity of stimulation signal. In the third experience, this result was used to complete the real-time muscle biofeedback system using LabVIEW on a PC. The result demonstrated that this system could adjust the intensity of stimulation according to the SEMG power. This feedback can control the SEMG power within a particular range and keep the muscle contraction stable. Although the proposed system accomplished the basic requirement, there are some uncertain factor influenced the computation of SEMG power. It is not sure that whether the muscle response arisen from higher frequency stimulation should be included in the bio-feedback. Additionally, a more efficient feedback system that can predict the SEMG power can be incorporated into the system in the future and the performance and stability of this system will be increased.
工學院 > 醫學工程研究所