Unified Multimodal Synchronization with Energy and Entropy Anchors: A Cross-Validated Framework for Wearable Sensing

Temporal synchronization of multimodal data streams remains a fundamental challenge in wearable sensing applications, particularly those involving eye tracking alongside in-person and environmental data. This paper presents a novel cross-validated synchronization framework that combines complementary methods across audio and inertial measurement unit (IMU) data channels. Our approach leverages both global cross-correlation and energy-anchored detection for audio synchronization, providing mutual validation between methods. For IMU signals, we introduce an entropy-anchored technique that isolates high-information-density regions as synchronization points. The framework was developed to address synchronization challenges faced in a particular research study involving multiple recording devices, eye-tracking glasses, a GoPro camera, and smartphone-based IMU recordings. However, the proposed methods are generalizable to analyzing temporal relationships between human gaze, body movements, and environmental data in complex, multi-device research scenarios.