Design Of Wireless Multi-channel Electromyography-Detecting System Based On Compressed Sensing

Paralysis has always been one of the major diseases affecting human health.There are more than 20 million paralyzed patients in China,and the paralyzed patients in USA account for about 1.9% of the total population.The rehablition of limb motor function has been a research hotspot in the field of rehabilitation.The whole project of this thesis aims to design a wireless multi-channel EMG?electromyography?rehabilitation training system based on the theoretical basis of communication and FES?Functional Electrical Stimulation?and the existing research of the laboratory to help patients with limb movement training and rehabilitation,which makes contributions to the cause of rehabilitation and medical treatment.It has therefore certain practical significance,scientific significance and social value.This project undertakes the design of the EMG detecting system of the whole system.The main research objective is to find a suitable CS?Compressed Sensing?algorithm and design a software and hardware system for the EMG signal detecting system capable of accurate detection and wireless transmission.Following works and results have been arrived in this project:1.This project conducts a theoretical study on the EMG signal,and combines the physiology of the EMG signal to design the hardware and software framework of the EMG detection system,and investigates the selection of each hardware module.The hardware system can be divided into EMG signal detection and transmission hardware module and Android host computer receiving hardware module.The EMG signal detection and transmission module can be divided into three separate hardware modules: detection module,communication module and processing scheduling module.The software system is divided into two parts: the MCU software part and the Android software part.2.The project conducts theoretical research on the CS algorithm and focuses on the reconstruction algorithm of CS.It conducts theoretical research,modeling simulation and performance comparison on three types of reconstruction CS algorithm: greedy algorithm,approximate l0 norm reconstruction algorithm and convex optimization algorithm.The greedy algorithm has advantages for the reconstruction of low-dimensional simple data.The convex optimization algorithm is suitable for the reconstruction of complex data,but its reconstruction time performance is poor.The approximate l0 norm reconstruction algorithm achieves a relatively good coordination performance in both reconstruction accuracy and reconstruction time.3.The project verifies that CS algorithm can be used for the CS sampling and reduction of EMG collected in this project.In the comprehensive consideration of reconstruction precision and reconstruction time,SL0 algorithm is selected as the reconstruction algorithm and then optimized,and then the optimized algorithm effect is verified: by reconstructing 10 k data points,the single-point error of reconstruction is controlled within 10^-4 V,and the reconstruction time is 0.698 s.4.The project has designed and implemented the hardware circuit for the system.By designing the detection module,selecting the communication module and the processing scheduling module,connecting three modules,the board implementation of the EMG signal detection and transmission module is completed.5.The software system of the project is designed and implemented.On the MCU side,the system initialization and signal preprocessing,signal compression sensing sampling and Bluetooth communication are realized.On the Android side,bluetooth communication,CS reconstruction and UI design are realized.