Correlating Time Series Signals and Event Logs in Embedded Systems

Area of Science:

• Computer Science
• Embedded Systems Engineering

Background:

• Embedded systems generate diverse data, including time series signals and event logs.
• Correlating these data types is challenging due to potential clock drift between devices.
• Accurate correlation is crucial for understanding device performance and anomalies.

Purpose of the Study:

• To develop a robust method for correlating time series signals with event logs in embedded systems.
• To address the challenge of clock inconsistencies between data acquisition and monitored devices.
• To enable effective analysis of device operation and internal activities.

Main Methods:

• Classifying event logs using a novel similarity metric to determine their temporal distribution.
• Matching identified event log sequences with patterns in time series data.
• Employing iterative time offset adjustment for accurate correlation, utilizing object-oriented data models.

Main Results:

• The proposed algorithms successfully correlate event logs with time series signals, accounting for clock drift.
• The method was validated through power consumption analysis on a Holter device, demonstrating effectiveness.
• The approach proved universal and adaptable for various device optimization tasks.

Conclusions:

• The developed correlation technique provides a reliable solution for analyzing embedded system data.
• This method enhances the ability to diagnose device behavior and optimize performance.
• The universality of the approach suggests broad applicability across different embedded systems.