Scholarlink

<p>Wireless Sensor Networks (WSNs) are one of the basic technologies that can enable applications of the Internet of Things (IoT), which are characterized by distributed sensing, environmental monitoring, and real-time data communication. Battery power of sensor nodes is however limited, and results in susceptibility to abnormal network behavior, which have a significant impact on routing efficiency, reliability, and network lifetime. However, to overcome these challenges, this paper suggests a Hybrid Optimization-Based Deep Learning Framework (HO-DLF) for energy-efficient routing and anomaly detection in WSNs with IoT. A lightweight CNN-ShuffleNet model is used to realize intelligent anomaly detection and the framework combines Genetic Algorithm (GA), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Cuckoo Search Algorithm (CSA) to find optimal energy-aware multi-hop routing paths. A 300-node metropolitan-scale simulation environment (1000&#xa0;m × 1000&#xa0;m) was created and used to test the proposed framework with a centralized base station and 300 random nodes deployed. The initial energy of each sensor node was set to 2&#xa0;J, and the communication range was set to 100&#xa0;m, which were realistic parameters for monitoring applications in an urban environment. To evaluate the routing behaviour under different loads of communications, a heterogeneous deployment consisting of residential, industrial, park and traffic regions was taken in account for the simulation. The results from the simulations prove that HO-DLF is able to achieve an anomaly detection rate of 98%, decrease average energy consumption to 1.2&#xa0;J, enhance the packet delivery ratio to 92.4%, and prolong network lifetime to 220 days. Comparative study with LEACH, GA, ACO, and CSA also illustrates the energy efficiency, reliability of routing, and sustainability of the proposed approach. The results revealed the capability of HO-DLF as an effective and scalable approach for the deployment of an IoT-enabled WSN for real-time smart city applications that require to be secure and energy-efficient.</p>

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

Hybrid Optimization Based Deep Learning Framework for Energy-Efficient Routing and Anomaly Detection in IoT-Enabled Wireless Sensor Networks

  • 期刊论文

  • 发表日期：

    2026年7月29日

查看全文
  • M. Parameswari,
  • P. Nancy,
  • A. Anist

摘要

Wireless Sensor Networks (WSNs) are one of the basic technologies that can enable applications of the Internet of Things (IoT), which are characterized by distributed sensing, environmental monitoring, and real-time data communication. Battery power of sensor nodes is however limited, and results in susceptibility to abnormal network behavior, which have a significant impact on routing efficiency, reliability, and network lifetime. However, to overcome these challenges, this paper suggests a Hybrid Optimization-Based Deep Learning Framework (HO-DLF) for energy-efficient routing and anomaly detection in WSNs with IoT. A lightweight CNN-ShuffleNet model is used to realize intelligent anomaly detection and the framework combines Genetic Algorithm (GA), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Cuckoo Search Algorithm (CSA) to find optimal energy-aware multi-hop routing paths. A 300-node metropolitan-scale simulation environment (1000 m × 1000 m) was created and used to test the proposed framework with a centralized base station and 300 random nodes deployed. The initial energy of each sensor node was set to 2 J, and the communication range was set to 100 m, which were realistic parameters for monitoring applications in an urban environment. To evaluate the routing behaviour under different loads of communications, a heterogeneous deployment consisting of residential, industrial, park and traffic regions was taken in account for the simulation. The results from the simulations prove that HO-DLF is able to achieve an anomaly detection rate of 98%, decrease average energy consumption to 1.2 J, enhance the packet delivery ratio to 92.4%, and prolong network lifetime to 220 days. Comparative study with LEACH, GA, ACO, and CSA also illustrates the energy efficiency, reliability of routing, and sustainability of the proposed approach. The results revealed the capability of HO-DLF as an effective and scalable approach for the deployment of an IoT-enabled WSN for real-time smart city applications that require to be secure and energy-efficient.