Research On Energy Efficiency-aware Heart Rate Detection Technology For Head Mounting

Keeping employees safe is one of the top responsibilities of every industry.As a kind of safety equipment used in the construction industry,smart helmets are constantly being improved with the rapid development of wireless communication technology and sensing technology.The real-time heart rate of the staff can measure the health status of the human body.The heart rate detection system is built into the smart helmet when the wearer is fatigued from long-term construction or the heart is unwell,the smart helmet can give an alarm in time.Compared with other heart rate detection technologies,photoplethysmography(PPG)heart rate detection technology has gradually attracted attention due to its advantages of non-invasiveness and portability.Although heart rate estimation from PPG signals has been applied in many fields,there are still many challenges,such as the robustness of PPG heart rate detection and the energy consumption of PPG heart rate detection systems.Herein,this thesis discusses the application of energy efficiency-related PPG heart rate detection real-time system in head-mounted devices,and verifies the feasibility and effectiveness of this technology through experiments.The main work of this thesis follows:(1)Aiming at the reliability of the collected signals,this thesis designs and builds a headmounted PPG signal collection system,including the design and implementation of the hardware system and software programs.The reliability of the system’s signal acquisition is verified by experiments.The experimental analysis is carried out on the best signal collection position of the forehead,and the results show that the signal quality collected in the middle of the forehead is the best.In addition,the signal is collected at the best collection position,the PPG and acceleration data sets are established,and the data are preprocessed.(2)Aiming at the robustness and time complexity of the PPG heart rate detection algorithm,this thesis proposes an adaptive heart rate extraction algorithm based on activity type.The types of head activity are divided into three types: stationary,non-periodic,and pseudo-periodic.First,the characteristics of the acceleration data are extracted,and the dimensionality of the extracted features is reduced by using linear discriminant analysis,and then the logistic regression classification model is used.The extracted features are trained,and finally an adaptive heart rate extraction algorithm is executed based on the model prediction results.The experimental results show that in the multi-scene activities designed in this thesis,compared with other single heart rate detection methods,the adaptive heart rate extraction algorithm based on the activity type proposed in this thesis can not only adapt to a variety of activity scenes,but also has better robustness,and the time complexity of the algorithm is lower.(3)Aiming at the energy consumption problem of head-mounted heart rate detection equipment,this thesis proposes an energy efficiency optimization strategy on the STM32F405 microcontroller and implements an adaptive heart rate extraction algorithm based on activity type.The energy efficiency optimization strategy proposed in this thesis includes a sleep wake-up mechanism and an automatic standby wake-up function.In the data reading stage,the current consumption of the running mode is 53.6m A,and the average duration is 0.8ms,and the current consumption of the sleep mode is 33.9m A,and the average duration is 9.2ms.When the user removes the device,the system automatically enters standby mode with a current consumption of 0.62 m A.Automatically wakes up into run mode when the user picks up the device.The energy consumption experiment results show that the energy consumption optimization strategy proposed in this thesis works well.In addition,this thesis implements an adaptive heart rate extraction algorithm based on activity type on the STM32F405 microcontroller.The mean absolute error(MAE)of heart rate detection under static,aperiodic activity and pseudo-periodic activity is 0.772,2.016,2.566,The pearson correlation coefficient(PCC)are 0.989,0.957,0.916 respectively,better than other heart rate detection methods.