The Controller Development Of The Prosthetic Hand

Prosthetic hand is a focus research filed in both robotics and biomedical engineering. The study on this area is of great significance to improve the living standards and to advance the development of the medical welfare for amputees. This dissertation studied the control system of the under-actuated humanoid prosthetic hand with 3 degree of freedom. After analyzing the structure of the prosthetic hand, we have designed both the control system and the EMG electrode circuit.Firstly, this dissertation summarized the status in quo of the prosthetic hand'development, analyzing the characteristics of all kinds of prosthetic hands and forecasting the research direction of it. Then we generally described the fundamental structure and basic functions of each subsystem in the prosthetic hand including under-actuated mechanism,the torque sensory system,the micro controller and the step motor diver system.Secondly, we discussed the integrated hardware system and the software system of the prosthetic hand in details. Taking TMS320F2812 DSP as its micro-controller, the hardware system comprises a specified programmable step motor controller and three big-currency micro-step drivers. This system can control 3 step motors at the same time,sample the torque and EMG signals and communicate with PC. Meanwhile, a strain-based,changeable and reliable torque sensor has been developed. The system software is modularized into 4 parts: DSP system configuration,step motor controlling and driving,A/D torque sampling and SCI communicationThirdly, we conducted the designing of the EMG electrode circuit.. In this dissertation, the EMG electrode circuit integrates the first differentiate amplifier,lowpass,50Hz power frequency trapping and the second amplifier. The EMG circuit has been simulated by using SPICE, furthermore, we have manufactured the EMG circuit board and finished some tests with it, the test results were satisfactory. Then we have do some research on the Parametric Autoregressive (AR) model. Lastly, we fulfilled all the subsystems'and the entire system's debugging. The open-loop position control experiment has been conducted. EMG signal control experiments are accomplished in the prosthetic hand system by analyzing the Parametric Autoregressive (AR) model. The identification rates of the thumb,index finger and middle finger are satisfactory, thus the prosthetic hand can grasp different shape objects through the EMG control.