Multimodal data synchronization: a high-level software methodology for heterogeneous devices

Area of Science:

• Biomedical Engineering
• Data Acquisition Systems
• Signal Processing

Background:

• Multimodal data acquisition from heterogeneous devices often suffers from synchronization errors due to clock skew and offset.
• Existing solutions typically require complex external hardware or shared reference signals, hindering deployment and scalability.

Purpose of the Study:

• To introduce a high-level software methodology for precise multimodal data synchronization.
• To eliminate the need for external synchronization hardware in data acquisition systems.

Main Methods:

• A dedicated data acquisition protocol records device-specific timestamps and a master PC timestamp.
• Offline linear regression models internal timestamps to a common reference, correcting clock discrepancies.
• Validation involved electrocardiographic (ECG) and Inertial Measurement Unit (IMU) devices against a gold standard.

Main Results:

• Achieved an average R-peak synchronization delay of 20.1 ms over a one-hour acquisition (0.03% difference).
• Demonstrated successful elimination of external synchronization hardware requirements.
• Scalability is limited only by PC connectivity.

Conclusions:

• The methodology offers a robust and scalable foundation for precise multimodal synchronization, contingent on device Software Development Kits.
• High-level software control can achieve high accuracy for dynamic clock adjustment in data acquisition.