Burst Delineation Using Error Metrics for Optical Burst Detection

The patent divides the bit stream into multiple subsets (first subset, second subset, etc.) and computes error statistics separately for each subset. This segmentation allows the system to capture different aspects of error patterns (e.g., individual bit errors vs. consecutive bit errors) without requiring a single complex error analysis mechanism, thereby improving delimiter detection accuracy while managing computational complexity through modular processing.

The patent applies different coefficients to different error statistics based on their specific characteristics and importance for delimiter detection. Rather than uniformly processing all error data with the same computational resources, the system assigns weighted coefficients (e.g., A for first error statistic, B for second error statistic) that reflect the local importance of each error type, thereby improving detection reliability while optimizing energy consumption by focusing computational effort where it matters most.

The patent combines multiple error statistics into unified error metrics through linear combinations (e.g., A*N1+B*N2). This merging process integrates information from different error analyses (individual errors, consecutive errors, overlapping subsets) into composite metrics that provide more precise delimiter detection. The combination approach maintains manageable complexity by using systematic linear formulations rather than requiring entirely new complex analysis structures.