Design Of Identity Authentication System Based On ECG And PPG Signals

In recent years,identity authentication has played an important role in people’s daily life.How to accurately,quickly and efficiently prove one’s identity to third-party equipment in order to obtain a certain service or enter a physical place is a key issue that need to be addressed in the information age.Electrocardio Gram and photoplethysmography signals are inherent physiological signals of the human body,which contain a large amount of physiological and pathological information.The Electrocardio Gram signal(ECG)and photoplethysmography signal(PPG)of different individuals are very different,which have good specificity and confidentiality.And it can be used as a new biological feature for identity verification.In this paper,the electrocardiogram and pulse signals generated by heart motion were the research objects.Through the analysis of the two signal characteristics,a wearable continuous identity authentication system based on the electrocardiogram and pulse signals was designed.The compressed sensing technology was combined to reduce the transmission volume of system data.While meeting the low power consumption requirements of wearable devices,the function of individual identity authentication was realized based on ECG and pulse body surface signals.The main research contents of this paper are as follows:(1)A low-cost,high-efficiency wearable heart signal measurement device was designed,which can realize the uninterrupted collection of the human body’s ECG and PPG signals,and provide a signal source for subsequent identification.The system mainly includes a microcontroller,ECG and PPG signal acquisition,analog-to-digital conversion,power supply,wireless transmission and other modules.The hardware circuit of the system is small in size,has excellent performance,and can realize accurate signal measurement.(2)Based on the research of compressed sensing theory,a sparse and reconstruction method for ECG and pulse signals was designed,and the K-SVD dictionary learning method was used to sparsely represent the two bioelectric signals.An improved Compressive Sampling Matching Pursuit(Co Sa MP)algorithm was proposed.The new algorithm overcame the shortcomings of unknown sparsity,which set variable step lengths to adjust the number of atoms,and shortened the signal reconstruction time.The improved algorithm was used to simulate the reconstruction of two kinds of bioelectric signals,and the evaluation parameters were introduced to analyze and compare their reconstruction performance.(3)The feasibility of an individual identity authentication model was analyzed based on ECG and PPG signals.According to the unique characteristics of the two signal waveforms,their own preprocessing and feature extraction schemes were designed,and stable and accurate signal characteristics were obtained through this scheme.In the authentication stage,the extracted signal features were identified and classified by the support vector machine,and self-collected cardiac signal data were used for classification experiments to verify the applicability and reliability of the designed scheme.The experimental results show that the identity recognition based on the two kinds of signals has reached a higher accuracy rate,the average recognition performance reached 94%.