A Study On The Control Of Prosthesis Using SEMG

Since myoelectric control is direct and natural, the prosthesis control by SEMGhas been the dominant sort in the exterior powered prosthesis. But there are also someproblems in myoelectric controlled prosthesis. For example, the user's muscle isprone to fatigue and the repeatability of the information pattern is poor. Furthermore,the influences of the exterior electric field are severe.In this thesis, the general features of Surface Electromyogram (SEMG)acquisition used as control signals of prosthesis are discussed. The requirements to thequalities of preeding circuits acquiring SEMG are presented. A high performancepre-amplifier and a filter are constructed. A SEMG acquisition system usingDAQ2006 and LabVIEW has been built. To evaluate the circuit, a group ofexperiments have been carried out. Some problems of the circuit are discussed basedon a comparison with commercially available NORAXOM MYOSYSTEM 1200.The experiments have proved that SEMG can be considered as a stationaryGaussian Process with zero mean and a controllable variance σ 2 and acorresponding mathematical model can be established.Human's hand is complicated and deft. In order to validate the myoelectriccontrol function, a simple artificial hand has been designed based on consulting thecommercial product.Threshold control is a simple method in myoelectric control. It works reliably. Anew scheme of control and drive circuit has been constructed using new IC. It is muchsmaller, consumes less power and has higher interference standing ability against thecommonly used circuits. Moreover, it is possible to reduce the cost of the prosthesis.The experiments have proved that the circuit works stably and the malfunction ratio ofthe artificial hand is low.