Time Synchronization of Signal Data in Multiparameter Measurement System
Saana Vallius · Theseus (Ammattikorkeakoulujen) · 2019
The objective of this thesis was to establish an accurate time synchronization of two patient monitoring sensors to collect synchronized photoplethysmogram (PPG) and electrocardiogram (ECG) signal data. ECG data were collected with a dual-parameter respiration sensor and PPG data with a pulse oximeter, respectively. The synchronized data are intended to be used in the development of continuous noninvasive blood pressure (cNIBP) monitoring methods based on pulse transit time (PTT). The required accuracy for the time synchronization was 1 millisecond over 10 minutes of simultaneous data collection. The sensors used in this project utilized a data ready interrupt feature that was used as a basis for the proposed time synchronization method. When using an interrupt line between the sensors it was possible to tie the data sampling events of the pulse oximetry sensor to the time base of the respiration sensor. The accuracy of the implemented time synchronization method was tested by introducing a disturbance to the data signals and observing the difference in the manifestation of that disturbance in the ECG and PPG data. Two 2-hour sets of data were collected by using two different disturbance signals, after which they were analyzed for both accuracy and precision. The precision tests were conducted by dividing the data to 10-minute batches and observing whether the accuracy of the time synchronization would change over time. The results showed that the maximum achievable accuracy of the proposed time synchronization method was 1.67, but the chosen validation method was only able to detect the accuracy within 9 ms. Furthermore, the precision tests showed no decrease in the accuracy over time. The switch mechanism used for starting the data collection produced a constant time offset in the data which had to be removed during post-processing. All obtained results show that it is possible to use the two sensors to collect accurately aligned PPG and ECG data. Although the 1 ms target was not met, the proposed method seems to be promising and can be assumed to maintain the accuracy for over the observed two hours. More work is needed in creating a more robust triggering mechanism for the simultaneous data collection and in creating a more accurate validation method.