Information-Theoretic Performance Analysis of Sensor Networks via Markov Modeling of Time Series Data

Area of Science:

• Information Theory
• Sensor Networks
• Signal Processing

Background:

• Passive sensor networks are crucial for detecting moving targets.
• Data-level information fusion is key for enhancing network performance.
• Quantifying individual sensor contributions is challenging.

Purpose of the Study:

• To develop an information-theoretic framework for analyzing passive sensor network performance in target detection.
• To formulate a novel measure for quantifying sensor information contribution.
• To validate the proposed method using experimental data.

Main Methods:

• Formulated a measure of sensor information contribution using symbolic dynamics.
• Approximated network information state using principal component analysis.
• Constructed Markov and x-Markov machine models to quantify sensor contributions via conditional entropy differences.

Main Results:

• The proposed method accurately identifies targets with very low false-alarm rates.
• Network decisions are independent of individual sensor behavior and identity.
• The algorithm demonstrated superior performance compared to a feature-level information fusion approach.

Conclusions:

• The developed information-theoretic method provides an effective approach for passive sensor network target detection.
• The sensor contribution measure offers valuable insights into network data fusion.
• The algorithm's independence from individual sensor characteristics enhances its practical applicability.