Network Time Synchronization Using Adaptive Window and Exponential Weighting

The patent applies dynamics by making the window size adaptive rather than fixed. The window size dynamically adjusts based on the number of propagation time measurements collected, transitioning from a small initial size to a larger size as more measurements are gathered. This dynamic adjustment allows the system to balance between computational simplicity and measurement accuracy, resolving the contradiction by adapting the window size to the current state of data collection.

The patent changes the parameter of window size from a static value to a variable that evolves over time. By gradually increasing the window size as more propagation time measurements are collected, the system optimizes the balance between computational load and measurement precision. This parameter change enables the system to start with low computational complexity and progressively improve measurement accuracy.

The patent implements periodic action by collecting propagation time measurements at regular intervals and processing them in batches. Instead of using each measurement immediately, the system accumulates multiple measurements over time and then computes the average when the window is full. This periodic batching approach smooths out jitter and wander effects while maintaining regular synchronization updates.

The patent applies preliminary action by pre-collecting multiple propagation time measurements before performing the final averaging computation. The system gathers measurements in advance, storing them in a window buffer, and only after collecting the required number of measurements does it compute the average propagation time. This preliminary data collection phase allows the system to mitigate jitter and wander effects before the actual synchronization calculation.

The patent uses dynamics by implementing a gradually increasing window size rather than using a large fixed window from the start. The window size grows as more measurements are collected, allowing the system to achieve high measurement accuracy over time without incurring excessive initial delays. This dynamic growth approach balances the trade-off between accuracy and time delay.

The patent applies preliminary action by pre-collecting measurements in an expanding window before final computation. The system prepares measurement data in advance with a growing window size, so that when synchronization is needed, the averaging computation can be performed quickly on already-collected data. This reduces the actual synchronization delay while maintaining the benefits of a larger measurement window.