An enhanced time synchronization method for a network based on Kalman filtering.

Area of Science:

• Computer Networking
• Distributed Systems
• Real-Time Systems

Background:

• Accurate time synchronization is critical for distributed systems and network protocols like IEEE 1588 (Precision Time Protocol).
• Existing methods face challenges with network latency, jitter, and clock drift, limiting synchronization precision.
• There is a need for robust and efficient time synchronization algorithms applicable across various network types.

Purpose of the Study:

• To introduce an enhanced time synchronization algorithm distinct from IEEE 1588 (PTP).
• To improve the accuracy, stability, and efficiency of time synchronization in networked environments.
• To provide a solution that is easy to implement and requires minimal hardware resources.

Main Methods:

• A unique synchronous message packet structure with fixed length (10 bytes) and highest system priority for preemptive transmission.
• Application of a Kalman filtering model to mitigate noise interference (clock drift, network delay jitter, asymmetry).
• Inclusion of a continuous clock drift compensation mechanism for sustained accuracy.

Main Results:

• Simulation results demonstrate minimal time offset (±1 clock cycle) in symmetric links with an 80 MHz crystal oscillator.
• Maximum time offset within ±3 clock cycles observed in asymmetric links.
• Significant reduction in time offset, from several microseconds to less than 40 nanoseconds, compared to original PTP and Kalman filter-based methods.

Conclusions:

• The proposed algorithm offers a highly accurate and stable time synchronization solution.
• Its design facilitates easy implementation and low hardware requirements, making it versatile for diverse networks.
• This enhanced method significantly improves upon existing time synchronization techniques, achieving sub-40 ns accuracy.