基於即時處理與FPGA實現之神經義肢驅動系統

Abstract

在所有生物電氣活動分析方法之中，電生理是較為直觀有效，精確與可靠並且最有工程應用潛力的技術手段。 本論文使用即時演算法,從相拮抗肌肉之EMG訊號中，藉由動態堆積模型，推算出拮抗肌肉活動所對應的控制參數，透過將控制參數輸入擬相程式與姿態映射矩陣，轉換成各組關節電機的操控指令，並進而由三維模型進行運動模擬。 除此之外，本研究透過由FPGA設計的平行驅動數位IC與相應之功率驅動電路，實現了透過擬相指令對機械肢體系統的驅動，執行抓握動作。

Among the techniques of analyzing bio-electricity, Electrophysiological method is a more straightforward,effective, precise and promising one. In this thesis,we used Real-time Processing Algorithm to calculate control parameters from Agonist-Antagonist muscles. By applying control parameters into Pseudophasic Model and Guesture Mapping Matrix,we got driving commands of joint motors,and emulated these commands on a 3D Hand Model. Further more,By using power driving circuit and parallel driving digi-IC which designed by FPGA we were able to drive prosthesis by these Pseudophasic command.