低AoI多无人机物联网任务分配和轨迹规划
Low AoI Multi-UAV IoT Task Allocation and Trajectory Planning
周子轩 1李新凯 1张宏立1
作者信息
- 1. 新疆大学电气工程学院,乌鲁木齐 830047
- 折叠
摘要
用信息年龄(AoI)能够有效衡量数据的时效性和价值.为了解决应急通信中平均AoI最小化的问题,将无人机作为信息中继,提出了一种基于深度强化学习框架的任务分配和轨迹优化算法.首先,通过分析AoI最小化与无人机任务的关系,将问题分解为2个阶段求解;其次,采用k-means++聚类算法为无人机分配任务,以飞行距离、受灾情况及救援小组需求为标准,基于指针网络框架优化了无人机的轨迹;最后,设计了集中信息共享机制,节省了能耗和信息分发时间.实验结果表明,相较于传统方法,所提的优化算法在无人机应急救灾中能够显著降低AoI,有效缓解临时通信的压力.
Abstract
Age of information(AoI)provides an accurate measure of the value of data.In order to minimize average Aol in emergency Internet of things(IoT)communication,unmanned aerial vehicle(UAV)is introduced as an information relay,and a task assignment and trajectory optimization algorithm based on a deep reinforcement learning framework is proposed.Firstly,by analyzing the relationship between AoI minimization problem and the UAV,it is solved in two stages.Secondly,the k-meanis++clustering algorithm is used to assign tasks to the UAV,and the trajectory of the UAV is optimized in real-time through the pointer network according to the flight distance,disaster situation and rescue team needs.Finally,a centralized information-sharing mechanism is designed to save energy consumption and information distribution time.The experimental results show that compared with traditional methods,the proposed optimization algorithm can achieve a smaller AoI in the UAV emergency disaster relief,thus alleviating the temporary communication pressure caused by emergency disaster relief.
关键词
无人机辅助物联网/信息年龄/任务分配/路径规划/深度强化学习Key words
unmanned aerial vehicle-assisted Internet of things/age of information/task assignment/trajectory optimization/deep reinforcement learning引用本文复制引用
出版年
2024
NETL
NSTL
万方数据