| Due to the benefits of contactless and insensitive to environmental factors such as light and temperature,vital sign(i.e.,respiration and heartbeat)monitoring based on continuous wave(CW)Doppler radar has shown great potential in applications,such as long-term monitoring of high-risk population,home healthcare,disaster relief,security,military and so on.However,even if there is no obstruction between the CW Doppler radar and the tester,the monitoring technology still faces some difficulties and challenges.First,under the interference of respiration harmonics,it is difficult to achieve the average heart rate measurement error of less than 1% when the measurement distance is greater than or equal to 1 meter,but in order to extract the heart rate with high accuracy under the condition of closer to the real monitoring scene,it is necessary to achieve this goal.In addition,in the case of insufficient frequency resolution(that is,frequency resolution greater than 0.1 Hz),it is extremely difficult to achieve the average measurement error of fast heart rate below 2%,but in order to better analyze the heart rate variability,it is also necessary to achieve this goal.Further,it is more difficult to extract both respiration rate and heart rate under the influence of specific random body movement(RBM),which has become a bottleneck that needs to be solved urgently.The above problems are studied,and the main contributions of the dissertation are as follows:1.For the problem of the respiration harmonics interference,the influence of loworder harmonics of the respiration on heartbeat signals was analyzed first.Based on the results of the analysis,the matched filtering algorithm was improved and optimized utilizing the characteristics that the matched filter can maximize the signal-to-noise ratio of the output signal and the polynomial fitting can help to filter out low-frequency interference signals.When the measurement distance is up to 1.5 meters,the heart rate can be successfully extracted with an average error of less than 1%.2.For the problem of insufficient frequency resolution,the pros and cons of various related algorithms in the literature were analyzed in this dissertation.For the problem that the fast heart rate cannot be extracted with high precision in a short time window in the literature,the frequency interpolation algorithm was improved.Using this algorithm,when the length of the time window is only 3 seconds,the fast heart rate with an average error of less than 2% is extracted successfully.3.Due to the complexity of the RBM,this dissertation focused on the common problem of uniform RBM,and established the time domain model of the whole body movement signal.Based on this time-domain model and a single receiver,and using the improved adaptive noise cancellation algorithm in this dissertation,both respiration rate and heart rate can still be extracted when the body movement speed is up to 47.6mm/s,while the highest body movement speed reported in the literature based on a single receiver is 20 mm/s. |
