Motion-Dependent Averaging for Physiological Metric Estimating

The patent implements a dynamic averaging window size that automatically adjusts based on detected motion levels. During high motion periods, the window size decreases to provide faster response and reduce motion artifacts. During low motion periods, the window size increases to improve signal smoothing and measurement accuracy. This dynamic adaptation resolves the contradiction by making the processing complexity worthwhile only when needed for precision.

The system changes the temporal parameter (averaging window size) of the signal processing based on motion detection thresholds. When motion exceeds a threshold, the system switches to a shorter averaging window; when motion is below the threshold, it uses a longer averaging window. This parameter adaptation allows the system to optimize measurement precision across different activity states without requiring permanently complex processing.

The patent applies different processing qualities to different segments of the signal based on local motion conditions. During high-motion segments, aggressive filtering is applied to reject noise. During low-motion segments, minimal filtering is applied to preserve physiological information. This local adaptation of filtering strength resolves the contradiction by applying noise rejection only where needed without sacrificing overall signal information.

The filtering strength dynamically adapts to motion levels rather than using a fixed filter. The system modulates the degree of signal averaging based on real-time motion detection, allowing it to be more aggressive during motion and more conservative during rest. This dynamic approach preserves physiological information that would otherwise be filtered out by static filtering methods.