黃漢邦臺灣大學：機械工程學研究所林昆柏Lin, Kun-BoKun-BoLin2007-11-282018-06-282007-11-282018-06-282005http://ntur.lib.ntu.edu.tw//handle/246246/61332本文的主要目的在於將本實驗室所研發的人工義肢系統改善成一個可攜式的系統。為了達成高度整合的目的，我們引進了SOPC的技術，並且使用PDA操控人工義肢，此外，我們也成功的把肌電辨識系統嵌進PDA中。 在人工義肢的控制器方面，我們利用硬體描述語言Verilog撰寫一個多指節控制器，並且使用Nios為整個人工義肢系統的主控核心。在肌電辨識系統方面，我們提出了一個新的時間-尺度領域（time-scaling domain）特徵，攫取此特徵並且透過支持向量機（support vector machine）進行分類，辨識率可達93％。 在PDA人機介面方面，我們開發了一套使用者圖形介面，透過這個圖形介面，使用者可以完整運用機械手的所有功能，包括肌電辨識系統的訓練與測試，抓杯子、雞蛋等抓握動作，以及感測器的監控等功能。In the thesis, the NTU-hand prosthetic system developed in our laboratory was transferred to a portable system. To obtain a highly integrated system, the SOPC technology was applied. A PDA was adopted to communicate with the NTU-hand prosthetic system. Besides, the EMG discriminative system was built on the PDA. To control the NTU-hand prosthetic system, the multi-joint controller was written in Verilog. Besides, the Nios was used as the kernel of the control system. In the EMG discriminative system, a new time-scaling feature is extracted and feed forwarded to the SVM, a 93% classification rate can be achieved. In the PDA-based human machine interface, a user friendly GUI was developed. Monitor and control of the prosthetic hand, the training and testing of the EMG discriminative system can be performed via the GUI.摘要 I Abstract II Contents III List of Tables V List of Figures VI Chapter 1 Introduction 1 1.1 Motivation 1 1.2 Related Works 2 1.2.1 Literature Survey of Prosthetic Hands 2 1.2.2 Literature Survey of EMG Discrimination Systems 4 1.3 NTU-Hand Prosthetic System 5 1.3.1 NTU-Hand IV Prosthetics 5 1.3.2 PDA-Based Human-Machine Interface 7 1.3.3 PDA-Based EMG Discriminative System 8 1.3.4 FES System 9 1.4 Windows CE .Net and Embedded Visual C++ 10 1.4.1 Windows CE .Net 10 1.4.2 Embedded Visual C++ 13 1.5 Thesis Organization 15 1.6 Contributions 15 Chapter 2 Control System of the NTU-Hand IV 17 2.1 Introduction 17 2.2 Kinematics of the NTU-Hand IV 19 2.2.1 Direct Kinematics of the NTU-Hand IV 20 2.2.2 Inverse Kinematics of the NTU-Hand IV 24 2.3 Multi-Joint Control Module 26 2.3.1 The Architecture of a Single Joint Controller 26 2.3.2 Implementation of PWM Generator 27 2.3.3 Implementation of a Quadrature Decoder 29 2.3.4 Implementation of the PID controller 31 2.3.5 Integration of Each Sub-Module 34 2.3.6 Building a Multi-Joint Controller 35 2.4 Development of Graphic User Interface 36 Chapter 3 Development of the EMG Discriminative System 41 3.1 Introduction 41 3.2 Wavelet Transform 47 3.2.1 Introduction to Wavelet Analysis 47 3.2.2 Continuous Wavelet Transform (CWT) 48 3.2.3 Discrete Wavelet Transform (DWT) 52 3.3 Support Vector Machine (SVM) 54 3.3.1 Introduction 54 3.3.2 Multi-Class Support Vector Machine 60 3.4 Implementation of an EMG Classifier 64 3.4.1 Signal Collection 64 3.4.2 Signal Pre-Processing 66 3.4.3 Feature Selection 67 3.4.4 Classification 70 3.5 Stand-Alone FES System 70 3.5.1 Implementation of the Human Machine Interface 72 Chapter 4 Enhancement of Grasp Capability 76 4.1 Force Sensing Resistor 76 4.2 Hardware Architecture of Force Sensing System 77 4.3 Grasp Strategy of the NTU Prosthetic System 78 Chapter 5 Experiments and Results 83 5.1 Experiment of Results of EMG Classification 83 5.1.1 Wavelet-Based Power 83 5.1.2 Windowing Power 85 5.2 Grasp Experiments 87 Chapter 6 Conclusions 92 6.1 Conclusions 92 6.2 Future Works 93 References 94en-US人工義肢支持向量機小波轉換prosthetic handsupport vector machinewaveletthesis