Research On Human Biometric Sign Monitoring And Identity Authentication System

As society pays more and more attention to healthy life,wearable devices begin to integrate more and more physical sign monitoring functions.Long-term detection of health status can effectively help medical workers improve the accuracy of diagnosis and reduce the rate of misdiagnosis;coupled with the epidemic in recent years,people have paid more and more attention to their own physical signs,especially body temperature information.At present,there is a lack of wearable devices for long-term body temperature detection in the market.Most of the devices require manual measurement and have poor portability.In the post-epidemic era,wearing a mask has become a daily habit for everyone,so we thought of integrating temperature measurement equipment into the mask to monitor people’s body temperature for a long time.The design needs to consider three aspects of light weight,high accuracy and real-time performance.Try to use miniaturized components to control the volume of the equipment,perform temperature calibration under different ambient temperatures and distances to ensure accurate readings,and use Bluetooth to connect with the user’s mobile terminal to achieve real-time monitoring.No matter what kind of wearable device it is for,it needs to accurately identify the user in order to provide accurate health-related information to relevant departments or institutions and provide users with customized services.If an additional sensor device is used to identify the user’s identity,it is against the requirement of light weight,and it will also increase the power consumption of the system.Therefore,thesis directly extracts the user’s identity information from the obtained temperature data.,to realize the one-to-one correspondence between user body temperature data and identity.In terms of hardware design,thesis chooses miniaturized components to design a wearable temperature acquisition system.Considering the volume limit,Arduino Nano is selected as the control core of the whole system;the AMG8833 infrared sensor is used for respiratory temperature measurement,and calculate the average of its pixsels readings to reduce the reading error of the device itself;The laser ranging module is designed to obtain the distance between the sensor and the human body;the HC-06 Bluetooth module is selected to realize wireless transmission with the mobile terminal,and the mobile terminal monitors and records the user’s temperature data in real time;and the power supply circuit is designed to ensure the long-term stability of the system run.In order to achieve the accurate reading of the system,the BP neural network optimized by genetic algorithm is used in thesis to establish the temperature compensation model under different ambient temperature and different temperature measurement distance,so as to effectively ensure the accuracy of the system reading.Using the collected temperature data,design the identification algorithm framework,including using discrete wavelet transform for denoising,selecting the Multi Rocket model for feature extraction,and combining three types of signals(basic signal,firstorder derivative signal,and autocorrelation signal)as algorithm input.SVM,Ridge Classifier CV algorithms and Attention-GRU are used for authentication,and the influence of different factors on the performance of the algorithm designed in thesis is compared.The results show that the authentication algorithm designed in thesis has certain robustness.