Scholarlink

<p>The given paper proposes a deep reinforcement learning framework for unmanned aerial vehicles (UAVs) navigation to provide optimal communication for the users. It comprises gated recurrent unit enhanced with graph attention mechanism to make effective use of inter-UAV communication network for enhanced data retrieval and decisions. To further improve exploration and robustness, the proposed framework is trained using maximum-entropy reinforcement learning, enabling UAVs to learn stochastic policies. A heuristic reward function, solely based on local observations, is designed to optimize global performance in terms of coverage, fairness, and energy efficiency. Extensive simulations validate the effectiveness of the approach, showing superior scalability, adaptability, and communication efficiency over existing methods.</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

A deep reinforcement learning framework for UAV navigation and selection of relay paths

  • 期刊论文

  • 发表日期：

    2025年9月13日

查看全文
  • Sumati,
  • Nikhil Kumar Marriwala,
  • Ram Avtar Jaswal

摘要

The given paper proposes a deep reinforcement learning framework for unmanned aerial vehicles (UAVs) navigation to provide optimal communication for the users. It comprises gated recurrent unit enhanced with graph attention mechanism to make effective use of inter-UAV communication network for enhanced data retrieval and decisions. To further improve exploration and robustness, the proposed framework is trained using maximum-entropy reinforcement learning, enabling UAVs to learn stochastic policies. A heuristic reward function, solely based on local observations, is designed to optimize global performance in terms of coverage, fairness, and energy efficiency. Extensive simulations validate the effectiveness of the approach, showing superior scalability, adaptability, and communication efficiency over existing methods.