An Integrated Framework for Data Quality Fusion in Embedded Sensor Systems.

Area of Science:

• Embedded Systems
• Data Science
• Sensor Technology

Background:

• Connected embedded sensor systems generate vast amounts of data crucial for vital applications.
• Ensuring the quality of this sensor data is paramount for reliable system performance and decision-making.
• Existing methods often lack a unified approach to assess and represent overall data quality.

Purpose of the Study:

• To propose a framework for fusing sensor data streams and their associated data quality attributes.
• To develop a single, interpretable value representing the current underlying data quality.
• To validate the fusion framework using real-world datasets and diverse data quality issues.

Main Methods:

• Defining data quality attributes and establishing metrics for quantifiable quality assessment.
• Engineering fusion algorithms based on maximum likelihood estimation (MLE) and fuzzy logic.
• Utilizing domain knowledge and sensor measurements within the fusion process.

Main Results:

• The proposed framework successfully fuses data streams and quality attributes into a meaningful data quality indicator.
• Both MLE and fuzzy logic-based fusion approaches demonstrated capability in detecting data quality issues.
• Verification on MEMS accelerometer and Intel Lab datasets confirmed the framework's effectiveness.

Conclusions:

• The developed fusion framework provides an interpretable indicator for assessing data quality in embedded sensor systems.
• The proposed methods offer a robust solution for managing and understanding data quality in complex sensor networks.
• This work contributes to enhancing the reliability and trustworthiness of data derived from connected devices.