부산대학교 도서관

Synchronization of signals in the post-trial analysis is a laborious process that is often a bottleneck during the signal analysis. As the concept of the Internet-of-Things (IoT) emerges and more sensors are implemented in a research trial, reliable synchronization schemes are becoming increasingly important. This article presents a synchronization algorithm that could align signals recorded by different platforms with different sampling frequencies to millisecond-level precision. The algorithm could also realign a recording that has been restarted after a device failure with the same alignment precision as the rest of the signals. The algorithm generates a secondary signal by permutation of six different step voltages in each cycle to produce a unique pattern before returning to a baseline. The algorithm has been deployed in an actual clinical trial involving 26 heart failure patients and five different bio-signal modalities. It has successfully aligned all trials, including one trial that had a device failure during the recording. Two aligned heart signals had an average beat-to-beat interval difference of 0.81 ± 0.79 ms or 0.90 ± 0.87% with no sign of a negative effect of the synchronization algorithm.