Detection And Recognition System Based On Neural Electrophysiological Signal

In recent years,biomedical technology continues to evolve,human-computer interaction has become more and more important,making human-computer interfaces a technology that has attracted much attention from researchers.The human-machine interface can collect the physiological signals of the human body,analyze the information related to the physiological changes of the human body,such as the EMG signal when the muscle is contracted or the brain electric signal when the attention state of the brain changes,and send this information to the computer.So as to realize the research of human physiological information.Under the needs of disabled patients,human-machine interfaces have developed rapidly in rehabilitation medicine.For example,for severely paralyzed people,the use of brain imagination to control robotic hands to assist life.Therefore,the design of an acquisition system based on neural electrophysiological signals not only has high academic research value,but also has real practical application value.In this paper,a wearable multi-modal brain signal attention detection system and a surface EMG control system are designed based on the multi-modal analog front-end chip independently developed by the research group.The main research contents are as follows:(1)The analog front-end chip independently developed by the test laboratory.The chip supports 8 input channels,3 different input modes,and multiple electrophysiological signal detection functions.In the DC test project,the power consumption of the chip is tested.The current consumption of each channel is about 50?A.Under the AC test project,the chip's frequency response bandwidth,gain,output noise,equivalent input noise converted back to the input,and in-band noise amplitude were tested,and compared with other chips on sale.The test showed that the independent research and development Availability of the chip.(2)Multi-modal brain signal attention detection system design: On the hardware,the EEG signal and cerebral blood oxygen signal acquisition module,the main control module,and the power management module are designed,and these modules are all designed with integrated chips.On the embedded software,the signal acquisition driver,BLE data transmission program,and low power consumption program design are designed.Considering the needs of low power consumption,the duty cycle of LEDs in the cerebral blood oxygen acquisition program is optimized,and the connection interval,number of packets,and packet length in the BLE data transmission program are uniformly debugged.On the host computer,it is designed based on python to display,store,process and recognize two kinds of physiological signals;in signal processing,independent component analysis is used to process brain electrical signals and Hilbert yellow transform to process cerebral blood oxygen signals;in recognition,transfer learning is used Recognition,the accuracy reaches 95%.(3)Surface electromyography control system design: Two control systems are designed,One is the robot control system.In terms of hardware,the EMG acquisition module,the main control module and the power management module are designed,in which the EMG acquisition module only uses 4 channels;on the embedded software,the EMG signal acquisition and transmission program is designed;on the upper computer,python is also used,and the linear discriminant analysis method is used to control the robotic hand.The other is a gesture recognition system,with 21 gestures designed.Different from the robotic hand control system,on the hardware,gesture recognition uses 8 EMG acquisition channels,which can obtain more muscle change information,and a separate surface EMG acquisition armband is designed to conform to the distribution of arm muscles;on the host computer,The EMG signal is processed by the least square method,and the accuracy of gesture recognition is more than 95% through deep learning.