Scholarlink

Wireless Body Sensor Networks (WBSNs) face key challenges in balancing energy efficiency, routing reliability, and data privacy under dynamic body movements and strict resource constraints. Clustering and routing protocols usually operate separately, which reduces their performance in adaptive situations and causes a lack of integrated privacy protections for sensitive healthcare data. This paper introduces an improved routing framework that combines Deep Embedded Clustering (DEC) with an Adaptive Contextual Routing Protocol (ACRP) to enable energy-efficient, context-aware, and privacy-protected communication in WBSNs. The approach features a multi-stage system: (1) autoencoder-based learning of spatial-temporal node features, (2) attention-driven Graph Neural Network (GAT) for mapping local topology, (3) DEC for flexible cluster creation with balanced cluster head selection, and (4) ACRP using a Deep Q-Network routing agent with multi-objective path evaluation. The protection provided by this protocol is achieved through differential privacy (ε-differential privacy), where Laplace noise is added, along with random and quantum-mechanical processes, to ensure the probabilistic direction of transmitted data. Simulation results applied to real physiological datasets (PhysioNet and MHEALTH multi-sensor systems) showed that the DEC + ACRP framework achieved significant performance improvements. This led to a 25–30% increase in network lifespan compared to traditional protocols, and achieved energy efficiency of up to 92%. In addition, this algorithm showed a decrease in end-to-end latency to 0.15 s and also reduced privacy leakage, i.e., high-level encryption, to around 0.06. It also raised the packet delivery rate to 94% compared to traditional protocols and some modern protocols. The results showed a security resistance exceeding 93% against attacks, including brute-force attacks, and also showed more than 88% against eavesdropping threats. Therefore, it showed efficiency in protecting and performing data with limited resources, making it a reliable method for fast and secure transmission of patient data and is considered a good and scalable proposal for real-time health monitoring.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhanced Energy-Efficient and Privacy-Preserving Routing Framework for Wireless Body Sensor Networks Using Deep Embedded Clustering and Adaptive Context-Aware Protocols

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Tuqa Kareem Jebur,
  • Lourdes Peñalver,
  • Jaime Lloret,
  • Haider K. Hoomod

摘要

Wireless Body Sensor Networks (WBSNs) face key challenges in balancing energy efficiency, routing reliability, and data privacy under dynamic body movements and strict resource constraints. Clustering and routing protocols usually operate separately, which reduces their performance in adaptive situations and causes a lack of integrated privacy protections for sensitive healthcare data. This paper introduces an improved routing framework that combines Deep Embedded Clustering (DEC) with an Adaptive Contextual Routing Protocol (ACRP) to enable energy-efficient, context-aware, and privacy-protected communication in WBSNs. The approach features a multi-stage system: (1) autoencoder-based learning of spatial-temporal node features, (2) attention-driven Graph Neural Network (GAT) for mapping local topology, (3) DEC for flexible cluster creation with balanced cluster head selection, and (4) ACRP using a Deep Q-Network routing agent with multi-objective path evaluation. The protection provided by this protocol is achieved through differential privacy (ε-differential privacy), where Laplace noise is added, along with random and quantum-mechanical processes, to ensure the probabilistic direction of transmitted data. Simulation results applied to real physiological datasets (PhysioNet and MHEALTH multi-sensor systems) showed that the DEC + ACRP framework achieved significant performance improvements. This led to a 25–30% increase in network lifespan compared to traditional protocols, and achieved energy efficiency of up to 92%. In addition, this algorithm showed a decrease in end-to-end latency to 0.15 s and also reduced privacy leakage, i.e., high-level encryption, to around 0.06. It also raised the packet delivery rate to 94% compared to traditional protocols and some modern protocols. The results showed a security resistance exceeding 93% against attacks, including brute-force attacks, and also showed more than 88% against eavesdropping threats. Therefore, it showed efficiency in protecting and performing data with limited resources, making it a reliable method for fast and secure transmission of patient data and is considered a good and scalable proposal for real-time health monitoring.