| 摘要: |
| 针对在复杂应用场景下使用能耗受限的无人机(Unmanned Aerial Vehicle,UAV)进行无线通信的问题，提出了一种针对无人机航迹同时考虑无人机飞行动力学约束的优化模型。模型中同时考虑了信息传输量和能耗，并使用两者之比来定义能效，将能效最大化作为优化目标。由于该问题非凸且缺乏闭式表达式，因此首先对该优化问题进行离散化处理，然后使用强化学习方法进行求解，并对算法的复杂度和收敛性进行了分析。在给定的实验场景中，使用数值模拟的方法给出了能效最大、能耗最低、直线飞行和全局能效最大化(Global Energy Efficiency Maximization,GREEN) 飞行4种飞行模式的最优航迹。实验对比发现，能效最大化飞行相比其他飞行模式，能效提升了2~10倍。 |
| 关键词: 无人机（UAV） 航迹优化 强化学习 能效最大化 |
| DOI：10.20079/j.issn.1001-893x.241011001 |
|
| 基金项目:陕西省自然科学基础研究计划(2025JC-YBMS-790) |
|
| A UAV Trajectory Optimization Method by Energy Efficiency Maximization Based on Reinforcement Learning |
| SHAN Bowei,QU Jiayuan |
| (School of Information Engineering,Chang’an University,Xi’an 710018,China) |
| Abstract: |
| For the unmanned aerial vehicle(UAV) wireless communication problem with limited energy consumption in complex application scenarios,a trajectory optimization model with taking into account the flight dynamics constraints of UAVs is proposed.The model considers both information transmission throughput and energy consumption,and uses the ratio of the two to define energy efficiency,with maximizing energy efficiency as the optimization objective.Due to the non-convexity and lack of closed form expression of the problem,the optimization problem is first discretized and then solved using reinforcement learning methods.The complexity and convergence of the algorithm are analyzed.In the given experimental scenario,numerical simulation methods are used to determine the optimal flight paths for four flight modes,including maximum energy efficiency,minimum energy consumption,straight-line flight,and Global Energy Efficiency Maximization(GREEN) flight.Through experimental comparison,it is found that the energy-efficient flight mode improves energy efficiency by 2~10 times compared with other flight modes. |
| Key words: unmanned aerial vehicle(UAV) trajectory optimization reinforcement learning energy efficiency maximization |
