Enhancing DBSCAN clustering with fuzzy system to improve IoT-based WBAN performance

Area of Science:

• IoT-based healthcare
• Wireless Body Area Networks (WBANs)
• Networked systems

Background:

• WBANs face dynamic conditions and resource constraints impacting data clustering and energy management.
• Traditional clustering methods like DBSCAN struggle with WBANs' large node counts, limited energy, and diverse data.
• Suboptimal performance arises from static parameters in conventional clustering techniques.

Purpose of the Study:

• To introduce a novel clustering approach for WBANs that dynamically adapts to network conditions.
• To enhance the DBSCAN algorithm using a fuzzy system for real-time parameter optimization.
• To improve data clustering accuracy, energy efficiency, and network stability in resource-constrained WBANs.

Main Methods:

• Enhanced DBSCAN algorithm incorporating a fuzzy system.
• Dynamic optimization of DBSCAN parameters (Epsilon and MinPts) using real-time node speed and RSSI.
• Comparative simulation analysis against traditional clustering methods (Classical DBSCAN, PSO Clustering, LEACH, PEGASIS).

Main Results:

• The proposed fuzzy-enhanced DBSCAN achieved superior clustering accuracy and energy efficiency.
• Significant improvements in cluster stability were observed: 80% over Classical DBSCAN, 28.57% over PSO, 38.46% over LEACH, and 20% over PEGASIS.
• Enhanced network lifetime and cluster quality demonstrated the approach's effectiveness.

Conclusions:

• The fuzzy-enhanced DBSCAN offers a robust solution for data clustering in WBANs.
• Dynamic parameter adaptation is crucial for optimizing performance in resource-constrained IoT healthcare environments.
• This method presents a promising advancement for reliable real-time health monitoring systems.