Scholarlink

Unmanned Aerial Vehicles (UAVs) are increasingly used across various industries due to their versatility and efficiency, but their limited battery life poses significant challenges for prolonged operations. We propose a novel DRL-based framework to optimize the allocation of resources in UAVs, aiming to enhance energy efficiency and extend the operational time of UAVs. Our approach dynamically adjusts UAV flight paths and communication tasks based on real-time conditions, reduces energy consumption and improves mission outcomes. The extensive simulations show that our DRL-based method outperforms traditional static algorithms, leading to significant energy savings and more efficient UAV operations. Simulations reveal that our proposed UAV Task Area Optimization (UTAO) algorithm surpasses Double Deep Q-Learning Network (DDQN) and Genetic Algorithm (GA), which enhances remaining energy by 11.32% and 28.17% as UAV numbers grow, and by 14.88% and 35.78% with an increase in task area number.

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

Energy-Efficient Resource Allocation in UAV: A Reinforcement Learning Approach

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Qingtian Wang,
  • Beining Feng,
  • Xu Duan,
  • Jiaying Zong,
  • Siyu Chen

摘要

Unmanned Aerial Vehicles (UAVs) are increasingly used across various industries due to their versatility and efficiency, but their limited battery life poses significant challenges for prolonged operations. We propose a novel DRL-based framework to optimize the allocation of resources in UAVs, aiming to enhance energy efficiency and extend the operational time of UAVs. Our approach dynamically adjusts UAV flight paths and communication tasks based on real-time conditions, reduces energy consumption and improves mission outcomes. The extensive simulations show that our DRL-based method outperforms traditional static algorithms, leading to significant energy savings and more efficient UAV operations. Simulations reveal that our proposed UAV Task Area Optimization (UTAO) algorithm surpasses Double Deep Q-Learning Network (DDQN) and Genetic Algorithm (GA), which enhances remaining energy by 11.32% and 28.17% as UAV numbers grow, and by 14.88% and 35.78% with an increase in task area number.