移动通信2025，Vol.49Issue(3)：61-71,11.DOI:10.3969/j.issn.1006-1010.20250205-0001
面向低时延、低功耗的无人机任务卸载与轨迹优化策略研究
Research on Low-Latency,Low-Power Unmanned Aerial Vehicle Task Offloading and Trajectory Optimization Strategies
黄新梁 1林鹏 1刘艳 1张治中1
作者信息
- 1. 南京信息工程大学,江苏 南京 210044
- 折叠
摘要
Abstract
This paper studies a mobile edge computing system with multiple drones and multiple base stations.We aim to minimize the energy consumption of the drones and the system latency while considering load balancing among the drones.The optimization problem for the flight trajectory of the drones is jointly formulated as a non-convex optimization problem involving computation offloading decisions.Given the dynamic changes in task processing latency,drone power consumption,and load balancing due to varying system demands in real-world environments,the parameters of the optimization problem face uncertainties and dynamics over time.In this paper,we propose a model-agnostic meta-reinforcement learning algorithm(MAML-DDPG)to address this optimization problem,enabling the drones'decision-making to quickly adapt to the time-varying optimization goals.We evaluate the meta-learning capability of the algorithm through simulation experiments.Simulation results indicate that,compared to existing methods,the proposed approach can rapidly adapt to dynamic optimization objectives while effectively reducing both the energy consumption of the drones and the system latency while satisfying drone load balancing requirements.关键词
负载均衡/轨迹优化/计算卸载/元强化学习Key words
load balancing/trajectory optimization/computation offloading/deep reinforcement learning分类
电子信息工程引用本文复制引用
黄新梁,林鹏,刘艳,张治中..面向低时延、低功耗的无人机任务卸载与轨迹优化策略研究[J].移动通信,2025,49(3):61-71,11.
