Swallowing Signal Analysis System Based On Simultaneous Measurement Of SEMG And SEIM

A balanced diet is essential to a healthy body,and it is necessary to monitor the daily food intake.This is not only conducive to the timely adjustment of the type and quantity of food intake,but also convenient for nutritional epidemiological research.However,the traditional diet monitoring methods mainly rely on the users' self-reports,which is neither convenient nor accurate.To solve the problem of automatic and accurate monitoring of food intake in a non-invasive way,a synchronous measurement method of swallowing signal based on electromyography(EMG)and electrical impedance myography(EIM)is proposed in this thesis,and the relationship between swallowing signal and food intake is established.First of all,the current research status of diet monitoring methods is described and the existing problems are pointed out.The relationship between swallowing and diet is clarified,and the types of swallowing monitoring techniques and the principles of measurement methods are outlined.An introduction to the swallowing process is made,then the changes and the interferences of EMG and EIM during swallowing process are analyzed,and the method of synchronous measurement of EMG and EIM is given.The common anti-interference methods and classification schemes are introduced.Secondly,the least mean square(LMS)adaptive filtering algorithm and the principle of adaptive notch filter are described in detail.Aiming at the interference problem in swallowing process,a variable step size LMS adaptive notch denoising method is proposed,and the method is simulated and applied in practice,which proves the superiority of the method in swallowing signal monitoring.The recurrent neural network is briefly introduced,but the gating principle of long short-term memory(LSTM)and gated recurrent unit(GRU)is emphasized,and the advantages of GRU neural network in processing swallowing signals are explained.Meanwhile,the GRU neural network classification and recognition model of swallowing signals is given to establish the relationship between swallowing signal and food intake.Then,a swallowing signal acquisition device based on synchronous measurement of EMG and EIM is designed,which can simultaneously monitor the EMG and EIM signal using a pair of electrodes.The device is mainly divided into synchronous sampling analog front end,digital control part,data transmission part and power part.The modulation and demodulation principle of impedance signal in analog sampling front end and the realization method of synchronous sampling are introduced in detail.The embedded software development process and the upper computer development process of this device are provided.Finally,the actual test of volunteers proves the accuracy of the proposed method:the recognition accuracy based on single EMG is 82.7%,the recognition accuracy based on single EIM is 85.7%,and the recognition accuracy based on the combination of both is 95.6%.The research results are of great significance to the development and application of diet monitoring technology.