ProjectsProject Details

Synchronizing Wearable & Clinical Data ​for Predictive Health Analytics

Project ID: 7723-2-24
Year: 2024
Student/s: Yair Rejwan and Maxim Skliarevsky
Supervisor/s: Hadas Ofir

In this work, we developed a code solution to synchronize physiological data collected from multiple measurement devices. The process began by converting diverse data formats into Excel files, visualizing the recorded signals through graphs to analyze and understand the synchronization challenge. We then implemented a method to align signals from various devices and consolidate them into a single file, with each type of data organized in separate tabs and synchronized to a unified timeline.

To validate the implementation, we conducted an experiment involving synchronized, deliberate "interference" events across all devices. The expected outcome was consistent timing of these events in the resulting file. However, analysis revealed significant time discrepancies between recordings from different devices.

To further verify the accuracy of the code, we tested it with synthetic input and demonstrated that the code worked as intended. This proved that the observed timing discrepancies were due to inconsistencies in the way devices recorded data and not to flaws in the synchronization algorithm.

Finally, another experiment was performed in which we were able to verify the hypothesis and fix the code to compensate for these inconsistencies.

Poster for Synchronizing Wearable & Clinical Data ​for Predictive Health Analytics
Collaborators:
Logo of Technion Bio-Motion Lab Collaborator
Technion Bio-Motion Lab
Close menu