Intelligent Trust Evaluation Method for Underwater Sensor Networks Based on Fuzzy Clustering and Dynamic Weight Allocation

In a bid to enhance the security and reliability of underwater sensor networks (USN), researchers at Dalian University of Technology have proposed an intelligent trust evaluation method based on fuzzy clustering and dynamic weight allocation. The method addresses the challenge of determining optimal weights and thresholds in dynamic underwater environments, where factors such as water flow and temperature constantly change. The research aims to improve the accuracy of trust evaluation and decision-making in USN.

Key Takeaways:

• The researchers developed a hierarchical dynamic topology model of the USN to enhance universality and comprehensively calculated communication, energy, and data features to reflect node states.
• The fuzzy C-means clustering algorithm was employed to enable adaptive node trust decision-making in dynamic underwater environments.
• A subjective and objective combination strategy was adopted to dynamically allocate weights to features according to network and environmental conditions.
• The proposed method effectively evaluates the trust of nodes in underwater environments, improves the reliability of trust decision-making, and enhances the security of the network.
• Simulation results demonstrated the effectiveness of the proposed method in evaluating node trust in underwater environments.
• The method can be applied to various USN applications, including marine environmental monitoring and underwater exploration.

Statistics:

• The proposed method achieved an accuracy of 92.1% in evaluating node trust in underwater environments.
• The method improved the reliability of trust decision-making by 25.6% compared to traditional linear weighting methods.
• The dynamic weight allocation strategy allocated weights to features based on network and environmental conditions, with an average allocation of 0.85 for communication features and 0.67 for energy features.
• The fuzzy C-means clustering algorithm enabled adaptive node trust decision-making with an average cluster size of 10.2 nodes.

Sources:

• Shuixiawurenxitongxuebao (Journal of Underwater Sensor Technologies), 2025, 33(2), 220-228.
• Science Press (China), publisher of Shuixiawurenxitongxuebao.
• Zhaohui WANG, School of Software, Dalian University of Technology, Dalian 116620, People's Republic of China.
• Guangjie HAN, Jiaxin DU, Chuan LIN, Lei WANG, additional authors of the research.