Design And Realization Of Vehicle Health Monitoring System Based On ECG

Fatigue driving and heart disease emergencies have become major hidden dangers in driving safety,and the resulting traffic accidents also occur from time to time.Therefore,it is urgent to take measures to monitor the physiological condition of the driver in real time and improve the driving safety.At present,in the field of intelligent auxiliary driving,China lacks a vehicle-mounted health monitoring device that is easy to promote.In this paper,ECG acquisition module,data preprocessing and indicator analysis module,and communication and display modules are combined to form an on-board health monitoring system based on ECG.The signal acquisition module is mainly responsible for the extraction of ECG signals.The acquisition device consists of a steering wheel and electrode pads;The signal processin g and analysis module mainly includes the hardware processing and algorithm analysis of th e signal,and the hardware processing platform integrates the heart rate monitor,the microc ontroller and other components.In the algorithm part,the collected ECG signal is firstly su bjected to de-basis drifting,filtering and amplification and other preprocessing,and then the RR interval sequence is extracted.Then,in the time-frequency domain analysis of the RR interval sequence,more than a dozen indicators for data training can be obtained.Then,PC A(principal component analysis)is performed on multiple time-frequency domain indicators to obtain low-dimensional matrices and corresponding transformations matrix.Then K-mean s clustering is performed using low-dimensional matrices to classify the corresponding categ ories of different physiological states and finally obtain the analysis results;The communica tion and display module is mainly composed of CAN communication analysis tool and upp er computer.The CAN communication tool is used to connect the data bus of the hardware platform with the upper computer,and is responsible for transmitting the data read from the bus to the upper computer for display and broadcast.The upper computer adopts Labview’s integrated module for its own development.At present,it can realize the real-time wavefor m display of the collected ECG signals,as well as the icon prompts and voice broadcasts of the six physiological state recognition results.This paper proposes an interpolation method based on partial least squares multiple RRI deletion,the partial model is constructed by partial least squares(PLS),only when the R wave error is detected,the missing RRI was inserted through a local regression model and a measured RRI threshold was used to test the R-wave detection error.Experimental results show that compared with the traditional interpolation method in case of multiple R wave loss,the proposed method increases the extraction rate of key information in the frequency domain analysis of HRV by nearly 60%,and the method helps to improve related health monitoring services based on HRV analysis.