Scholarlink

In IoT systems, WSNs are another importance technology used in IoT systems for monitoring and controlling through real-time sensor data gathered in different area of interest like smart city, agricultural, health, etc. Nevertheless, energy-saving can still be an acute problem since sensor nodes are characterized by limited power supply and since reliable, low-latency data transfer is mandatory. As part of the solution proposal, this research introduces a novel optimization model that combines DRL and session-based dynamic resource management for wirelessly powered IoT (WPIoT) sensors. The framework under consideration utilizes the DDQN to implement the power control strategy and spectrum sharing in P2P communication while the Policy Gradient to manage Charging Transmission Control of WPIoT sensors. The simulation proves that the proposition of the hybrid of MPL, DRL, and Fuzzy logic reduces energy use by 27%, latency by 17 ms and has a 94% success rate for DMSS data transmission sessions above the baseline DRL. This work contributes to the state of art by giving an integrated solution that solves interference management, energy harvesting and scalability in heterogeneous WSNs.

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 Energy Optimization in Wireless Sensor Networks Using Integrated Deep Reinforcement Learning and Dynamic Resource Allocation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • S. Saraswathi,
  • R. Mahaveerakannan,
  • R. Balamanigandan

摘要

In IoT systems, WSNs are another importance technology used in IoT systems for monitoring and controlling through real-time sensor data gathered in different area of interest like smart city, agricultural, health, etc. Nevertheless, energy-saving can still be an acute problem since sensor nodes are characterized by limited power supply and since reliable, low-latency data transfer is mandatory. As part of the solution proposal, this research introduces a novel optimization model that combines DRL and session-based dynamic resource management for wirelessly powered IoT (WPIoT) sensors. The framework under consideration utilizes the DDQN to implement the power control strategy and spectrum sharing in P2P communication while the Policy Gradient to manage Charging Transmission Control of WPIoT sensors. The simulation proves that the proposition of the hybrid of MPL, DRL, and Fuzzy logic reduces energy use by 27%, latency by 17 ms and has a 94% success rate for DMSS data transmission sessions above the baseline DRL. This work contributes to the state of art by giving an integrated solution that solves interference management, energy harvesting and scalability in heterogeneous WSNs.