无人机集群编队控制方法研究综述  

A Review of UAV Cluster Formation Control Methods

作　　者：吴杰宏 李丹阳 WU Jiehong;LI Danyang(School of Computer Science,Shenyang Aerospace University,Shenyang 110136,China)

机构地区：[1]沈阳航空航天大学计算机学院,辽宁沈阳110136

出　　处：《无线电通信技术》2023年第4期589-596,共8页Radio Communications Technology

基　　金：国防基础科研项目(JH2021010)。

摘　　要：随着无人机技术的进一步发展,多无人机编队飞行的研究也受到了越来越多的关注。无人机相互配合组成编队群,可以充分发挥单个无人机所不具备的优势,更能胜任复杂、多任务场景下的工作。对无人机集群编队控制方法进行分类,分为传统控制法、群体智能算法、深度强化学习算法;对各类方法进行分析,着重归纳讨论了领导者-跟随者方法、人工势场法、运动学控制方法、蚁群优化算法、粒子群优化算法、人工蜂群算法、深度Q网络算法、深度确定性策略梯度算法、多智能体深度确定性策略梯度算法,并给出各自的优劣势;对无人机集群编队控制方法进行总结,指出传统控制法已接近成熟,但为了实现无人机的智能自主协同编队控制,仍需在群体智能算法和深度强化学习算法上融合新的思想与改进,从而发挥大数量无人机在复杂场景下的优势。With the further development of UAV technology,the study of multi-UAV formation flight has received more and more attention.UAVs cooperate with each other to form formation swarms,which can give full play to the advantages that individual UAVs do not possess and are more capable of working in complex,multi-task scenarios.UAV cluster formation control methods are classified into traditional control methods,swarm intelligence algorithms,and deep reinforcement learning algorithms.Various methods are analyzed,focusing on the leader-follower method,artificial potential field method,kinematic control method,ant colony optimization algorithm,particle swarm optimization algorithm,artificial bee colony algorithm,deep Q-network algorithm,deep deterministic policy gradient algorithm,and multi-agent deep deterministic policy gradient algorithm.Advantages and disadvantages of each are presented.UAV cluster formation control methods are summarize.It is concluded that traditional control method is close to maturity,but in order to achieve intelligent autonomous cooperative formation control of UAVs,it is still necessary to integrate new ideas and improvements in group intelligence algorithm and deep reinforcement learning algorithm,so as to take advantage of large number of UAVs in complex scenarios.

关 键 词：无人机集群 编队控制方法 深度强化学习 群智算法 路径规划 

分 类 号：V279[航空宇航科学与技术—飞行器设计] TN97[电子电信—信号与信息处理]