The Research Of WSN-based SEMG Driven Human-machine Interface Technology

sEMG (Surface Electromyography, sEMQ) produced by the body as an important biological signal can reflect the state of activities and muscle's function. It has been widely used in sports medicine, bionics, bio-feedback and rehabilitation projects. Wireless sensor networks (WSN) integrates sensors, computers, network communications, wireless transmission and embedded technology, forms a network through multiple nodes, inspects important information in network coverage in real time and allows users to remotely monitor network coverage area. This paper aims to combine the sEMG and WSN for human-machine interface technologies. We need to do a lot of research and finish the research of WSN-based sEMG driven human-machine interface technology.At first a sEMG signal amplification circuit was produced in this paper. Secondly, a wireless transmission experimental platform was constructed based on GAINS system that was based on ATmega128L, including collecting-data wireless node and sink node. We designed and implemented communication protocols of the network nodes' each layer and the serial communication between computer and sink node. At the same time the received sEMG signal data in computer was preprocessed and divided into motion pieces by using the improved method that was average energy difference. Then the feature information was extracted by statistical methods, AR parameter model method and the wavelet decomposition method. At last BP artificial neural network was designed and used to identify two kinds of actions, flexion of elbow and fist. And satisfied identification results were got, which laid the foundation for the subsequent research of developing the actual human-machine interface technology.