Interface Technology Research Based On Upper Limb’s Motion

Recently, the research of human machine interface technology has entered a period of rapid development in global disgrace. Development of the technology aims at making easier communication between people and machine, so that machine can replace people to do some work. It’s now involved in medical field, virtual reality field and so on. The movement of upper limb always reflects people’s intention, so upper limb information has been collected and analyzed.In the first part, this paper describes the characters of upper limb movement:surface EMG signal and acceleration signal. The overall structure of the interface system is mainly composed of four parts: electrical circuit of the surface EMG sensor, acceleration sensor circuit, signal transmission module based on ATmega8, processing module of computer classification.Then, the hardware circuit and software architecture are specified. The surface EMG signal circuit includes the design circuit, filter circuit, level converting circuit. The acceleration information circuit is based on acceleration sensor ADXL345. The software part focuses on the algorithm of SVM classification method which is used in the data analysis and discussed in detail in the end of chapter four.Using SVM algorithm, off-line analysis of EMG and acceleration information is realized. It includes the data dimensionality reduction, parameter optimization and classification of training model.The test results show that the accuracy of off-line data classification can achieve good results, finally, the SVM classification is ported to the computer to conduct online test. The results of experiments are consistent with the conclusion obtained by off-line test.