Timing Recovery Over Packet Networks Using Adaptive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Packet delay variations in packet switched networks lead to high jitter and wander in clock synchronization, making it challenging to maintain synchronization for legacy services over IP-based networks, especially when initial delay estimates are inaccurate or subject to outliers.
Innovation Solution
A minimum-statistics adaptive filter predicts delays to select timing packets closest to the predicted value, which are then used in a phase locked loop for robust frequency and phase synchronization, allowing for adaptive tracking and outlier resistance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If timing packets are filtered based on initial delay estimate, then clock synchronization accuracy is improved, but the method fails when initial estimate is inaccurate or affected by outliers
Solution Approach 1:
The filter adapts dynamically by adjusting the smoothing window size based on the observed stability of delay estimates. When delay variations are small, a larger window provides better filtering. When rapid changes occur or outliers are detected, the window size decreases to maintain tracking accuracy, allowing the system to respond to changing network conditions
Solution Approach 2:
The system continuously monitors the stability of delay estimates and uses this feedback to adjust the smoothing window size. This feedback mechanism allows the filter to automatically adapt to network conditions, maintaining robust performance without requiring precise initial delay estimates or being sensitive to outliers
2Stability of the object's composition
If a fixed smoothing window is used, then filtering stability is improved, but tracking behavior is limited when delay changes rapidly
Solution Approach 1:
The smoothing window size is made dynamic rather than fixed. The system automatically adjusts the window size based on the observed characteristics of delay variations in the network. This allows the filter to maintain stability under normal conditions while adapting quickly when delay changes occur, such as during network rerouting events
3Adaptability or versatility
If packet delay variations occur, then network flexibility is improved, but clock synchronization quality deteriorates due to high jitter and wander
Solution Approach 1:
An adaptive minimum statistics filter is introduced as an intermediary between the received timing packets and the clock recovery process. This filter mediates the effect of packet delay variations by selectively smoothing delays based on observed network conditions, reducing jitter and wander while preserving the benefits of packet network flexibility
Solution Approach 2:
The system changes the filtering parameter (smoothing window size) dynamically based on network conditions. By adjusting this parameter, the system optimizes the trade-off between filtering out jitter and wander versus maintaining responsiveness to legitimate delay changes, thereby improving clock synchronization quality without sacrificing network flexibility
Data Source
AI summary
In a method of recovering timing information over packet networks, raw network delays are measured using timing packets sent between a transmitter and receiver. The expected delay is predicted using a minimum statistics adaptive filter to track local minima of measured time delays over a smoothing window. Only those incoming timing packets which meet a particular criterion relative to the expected delay within a smoothing window are selected, and a local clock is adjusted based on the measured timing delays from the selected timing packets.