An Affinity Propagation-Based Self-Adaptive Clustering Method for Wireless Sensor Networks

Area of Science:

• Computer Science
• Network Engineering

Background:

• Wireless Sensor Networks (WSNs) are crucial for the Internet of Things (IoTs).
• Existing clustering methods for WSNs suffer from uneven cluster head distribution and unbalanced energy consumption.
• Machine learning offers potential solutions for intelligent WSN clustering.

Purpose of the Study:

• To present a novel Affinity Propagation-based Self-Adaptive (APSA) clustering method for WSNs.
• To enhance network performance by addressing drawbacks of traditional clustering techniques.
• To improve cluster head distribution and energy consumption balance in WSNs.

Main Methods:

• Combined K-medoids with Affinity Propagation (AP) for clustering.
• Utilized AP to determine the number of cluster heads and find optimal initial centers for K-medoids.
• Employed a modified K-medoids algorithm for iterative network topology formation.

Main Results:

• The proposed APSA method demonstrated superior performance compared to UCR-H, LEACH-AP, and EDDUCA algorithms.
• APSA effectively mitigates issues of homogeneous clustering and slow convergence rates associated with traditional K-medoids.
• Achieved more reasonable clustering performance in WSNs.

Conclusions:

• The APSA clustering method offers an effective solution for improving WSN performance.
• APSA enhances energy efficiency and load balancing in wireless sensor networks.
• This approach represents a significant advancement in intelligent WSN clustering.