Clock Drift Prediction Using Temperature-Aware ML Compensation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing cellular communication networks face challenges in maintaining accurate clock synchronization due to clock drift, which can lead to misalignments and loss of information, especially in applications requiring precise timing such as 5G, SDT, IIoT, and XR, where clock synchronization accuracy in the order of tens of picoseconds is necessary.

Innovation Solution

Implementing a machine-learning model, such as deep neural networks, to predict clock drift by modeling temperature and hardware aging effects, followed by a hybrid ML model for robust clock drift compensation, adjusting reported clock values to true clock times.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional clock synchronization methods are used, then basic timing is maintained, but clock drift causes misalignments and information loss in precision-critical applications

Engineering Contradiction:
Improveclock synchronization accuracyVSAvoiddata transmission reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent changes the parameters used for clock drift compensation by introducing temperature and aging as additional variables. The compensation parameter is dynamically adjusted based on measured temperature deviations and estimated aging effects, rather than using a fixed or simple adaptive algorithm. This allows the system to maintain picosecond-level synchronization accuracy under varying environmental conditions.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces traditional mechanical or simple electronic clock synchronization mechanisms with a software-based machine learning model. The neural network processes multiple input parameters (temperature, aging estimates, observed drift) to predict and compensate for clock drift, substituting physical synchronization hardware with an intelligent software system that can adapt to complex patterns.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Device complexity

If simple clock compensation methods are used, then basic timing is maintained, but temperature and aging effects cause accumulated drift

Engineering Contradiction:
Improvesynchronization system complexityVSAvoidclock drift prediction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements preliminary action by continuously measuring temperature and estimating aging effects before they cause significant clock drift. The machine learning model is trained in advance on historical data to learn the relationships between these parameters and clock drift patterns. This proactive approach allows the system to compensate for drift before it accumulates to problematic levels.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces temperature sensors and aging estimation mechanisms as intermediary elements between the physical clock hardware and the synchronization software. These intermediaries provide additional information about the clock's operating conditions, enabling more accurate drift prediction. The machine learning model acts as an intermediary that processes these multiple inputs to produce a refined compensation value.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If no temperature compensation is applied, then system simplicity is maintained, but environmental variations cause clock instability

Engineering Contradiction:
Improveclock synchronization reliabilityVSAvoidcompensation algorithm complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements dynamics by making the compensation algorithm adaptive rather than static. The machine learning model continuously learns from new data and adjusts its predictions based on changing environmental conditions and aging patterns. The compensation parameter is dynamically updated based on current temperature measurements and the model's learned relationships, allowing the system to respond to varying conditions optimally.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4199386B1Predicting clock drifting
Publication Date: 2025.11.19 NOKIA SOLUTIONS & NETWORKS OY
  • EP4199386B1 patent drawingFigure 1
  • EP4199386B1 patent drawingFigure 2
  • EP4199386B1 patent drawingFigure 3

AI summary

Disclosed is a method comprising obtaining a plurality of previous clock skews, a reported temperature and a reported time, based on the plurality of previous clock skews, the reported temperature and the reported time, obtaining a prediction of the current clock skew, determining a current clock offset based on the predicted current clock skew, determining a clock adjustment based on the current clock offset and the reported time, and determining a corrected time based on the clock adjustment.