An uncertainty-based distributed fault detection mechanism for wireless sensor networks.

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Wireless sensor networks (WSNs) face challenges with excessive message exchanges for fault detection, leading to degraded network quality of service and high energy consumption.
• Traditional distributed fault detection mechanisms suffer from misjudgments due to uncertainty and data loss.

Purpose of the Study:

• To propose an uncertainty-based distributed fault detection algorithm for WSNs that minimizes communication overhead and enhances detection accuracy.
• To address the impact of sensing measurement loss and uncertainty on fault detection performance.

Main Methods:

• Utilized Markov decision processes to effectively fill in missing sensing data, accounting for measurement loss.
• Employed evidence fusion rules based on information entropy theory and a degree of disagreement function to improve fault detection accuracy.
• Developed a distributed fault detection approach incorporating aided judgment from neighboring nodes.

Main Results:

• The proposed algorithm significantly reduces communication energy overhead compared to traditional methods.
• Demonstrated a higher ratio of accurate fault detection, mitigating misjudgments caused by uncertainty.
• Effectively handled missing sensing measurements through data imputation.

Conclusions:

• The uncertainty-based distributed fault detection method offers an effective solution for WSNs, balancing energy efficiency and detection accuracy.
• The integration of Markov decision processes and advanced evidence fusion rules enhances the robustness and reliability of fault detection in WSNs.