AF-DHNN: Fuzzy Clustering and Inference-Based Node Fault Diagnosis Method for Fire Detection

Area of Science:

• Computer Science
• Electrical Engineering
• Artificial Intelligence

Background:

• Wireless Sensor Networks (WSNs) are crucial for fire detection but face challenges in traditional node fault diagnosis.
• Existing methods suffer from complex structures, high computational demands, unstable data detection, and local minima issues.

Purpose of the Study:

• To propose a novel and efficient fault diagnosis mechanism for WSN nodes.
• To address the limitations of traditional WSN fault diagnosis methods.

Main Methods:

• Utilized fuzzy theory and an Adaptive Fuzzy Discrete Hopfield Neural Network (AF-DHNN).
• Extracted sensor status using filtered signal root mean square (FRMS) and normalized summation of positive amplitudes of the difference spectrum (NSDS).
• Implemented distributed fuzzy inference, Fuzzy C-Means Algorithm (FCMA), Sorting and Classification Algorithm, and an improved Discrete Hopfield Neural Network (DHNN).

Main Results:

• The proposed AF-DHNN method demonstrated prompt and effective diagnosis of abnormal WSN node faults.
• The system achieved associative memory and improved search efficiency through optimized sensor status information and diagnostic levels.
• Pre-alarming evident abnormal nodes' status saved diagnostic time.

Conclusions:

• The AF-DHNN offers a reliable and efficient solution for WSN node fault diagnosis in fire detection.
• This approach significantly enhances the overall reliability of Wireless Sensor Networks.