本論文主要在設計與實現一套肢體同動機械手臂系統來驗證以肌電訊號控制機械手臂的可行性，並作為穿戴式機械骨骼系統之先期研究。EMG(electromyography)訊號是肌肉在活動中收縮而產生的一種類比生理訊號，此訊號會因人體的意圖而產生不同的狀態，因此利用肌電訊號作為義肢或機械手臂的控制命令，是相當直接且自然的方法。這套系統以數位訊號處理器(TMS320F2812)實現肌電訊號的訊號處理、手部動作辨識、肌力估測方法與機械手臂運動控制(扭力控制與位置控制)，讓機械手臂可以實現受測者想要完成的動作和指令，達成機械手臂能跟隨受測者手臂動作的目的，也意味著未來應用於穿戴式機械骨骼系統時，機械手臂能幫助病人作其想要的動作，使生活更為便利，並提升其生活品質。 This thesis aims to develop an EMG-based upper-limb motion synchronization robot arm system which enables an upper-limb robot arm to move synchronously with the tester. EMG is a physiological signal generated during muscle contraction. With an appending in implicating the motion intention, he EMG is very suitable to be used for robot control and prosthesis control in an intuitive and directive manner. In this thesis, we combine four sections and realize algorithms with a digital signal processor (TMS320F2812). We use a digital signal processor (TMS320F2812) to measure the EMG signal, which imply the human subject’s intention. We feature extraction methods convert EMG signals to standard deviation (DEV) which is random variable. Then we design a system which can recognize human’s arm action ,and employ an estimator to estimate the muscle strength of the tester’s forearm. Therefore, the estimated muscle strength can be used as a torque command for robot arm system. We also use digital signal processor to realize the torque control and position control of robot arm system. So the robot arm can achieve the desired motion of the tester. In the future, this system can be transferred into an exoskeletal robot for human motion support, especially for those physically weak ones.
