Scholarlink

In the emergency communication system, unmanned aerial vehicles (UAVs) are often used as aerial base stations that provide disaster recovery communication services for ground users in disaster areas. The deployment efficiency (speed and accuracy) of UAVs is closely related to the efficiency of rescue, and efficient deployment can greatly improve rescue efficiency. To address this problem, we use Integrated Sensing and Communications (ISAC) to improve the accuracy of communication services provided by UAVs and use a Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm to accelerate the deployment of UAVs. Based on the MADDPG algorithm, the position selection of UAV agents in each time slot is determined by predicting user movements and selecting optimal actions from predefined action sets. The UAVs are deployed to the specified location with the goal of optimizing the achievable communication rate for users while adhering to the motion constraints and target perception performance requirements of the drones. Finally, the simulation results show that the proposed algorithm have effective performance on achievable rate.

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 Approach to UAV Path Planning for Integrated Sensing and Communications in Emergency Communication Environments

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yuxin Wu,
  • Songlin Sun,
  • Chenwei Wang

摘要

In the emergency communication system, unmanned aerial vehicles (UAVs) are often used as aerial base stations that provide disaster recovery communication services for ground users in disaster areas. The deployment efficiency (speed and accuracy) of UAVs is closely related to the efficiency of rescue, and efficient deployment can greatly improve rescue efficiency. To address this problem, we use Integrated Sensing and Communications (ISAC) to improve the accuracy of communication services provided by UAVs and use a Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm to accelerate the deployment of UAVs. Based on the MADDPG algorithm, the position selection of UAV agents in each time slot is determined by predicting user movements and selecting optimal actions from predefined action sets. The UAVs are deployed to the specified location with the goal of optimizing the achievable communication rate for users while adhering to the motion constraints and target perception performance requirements of the drones. Finally, the simulation results show that the proposed algorithm have effective performance on achievable rate.