Analysis And Application Research Of Pulse And Motion Signals On Wearable Electronic Eyeglass

Currently, cardiovascular diseases and chronic sports injuries have become a major cause of disability and death in humans, which has a serious impact on people’s quality of life and health. With the advantages such as portability, protection of user privacy, wearable detection technology has been applied to monitor human physiological signals and motion information.On the base of wearable eyeglasses system, different algorithms of physiological signal analysis and processing are designed. The main works include:Based on the analysis of pulse signal, the pulse signal pre-processing algorithms include using band-pass filter to filter pulse signal, applying Least Mean Square algorithm (LMS) to remove motion artifacts, and adopting cubic spline interpolation to remove the baseline drift.The pulse rate calculation algorithm based on the improved maximum method includes detecting the main wave peaks and moving average method. Firstly, track the maximums, set the method of maximum average threshold and the minimum time difference between the adjacent peaks; Then find the first main wave peak according to method of maximum average threshold, and determine the subsequent main wave peaks by minimum time difference and method of maximum average threshold. When the maximum isn’t the main wave peak, the maximum average is updated. Finally, use the main wave peaks to calculate pulse rate, and perform moving average method on all pulse rate results.Based on the variance threshold method, the method identifying daily actions includes the variance calculation, the threshold settings, and the weighted judgment. Firstly, calculate the variances of before-after and vertical acceleration signals. Then set the variance thresholds under the standing, slow walking, fast walking and jogging actions, and obtain each recognition results of the before-after and vertical directions. Finally, obtain the final action identification result by setting the weighted coefficients of above acceleration directions.Based on improved peak detection algorithm, the effective steps detecting algorithm includes weighted filtering, peaks judgment and steps detection. Firstly, the weighted filter is applied to original acceleration signals. Then, set the correction factor to judge the signal peaks. Finally, judge the first step by the original signal difference, and judge other steps by the time difference range from 0.2s to 2.0s between the adjacent steps.To verify the validity of the algorithms, a large number of experiments are designed to apply the algorithms to process the collected pulse and acceleration data on MATLAB platform. Experimental results show that the algorithm can achieve the functions of denoising of pulse signals, pulse rate calculation action identification and steps calculation.