With the increasing research interest and application demand for unmanned aerial vehicles in both civil and military fields,traditional model-based network deployment,design,and operation methods can hardly cope with dynamically changing unmanned aerial vehicles scenarios.This paper reviews AI-based intelligent networking technology,which offers high flexibility and adaptability and introduces reinforcement learning as an important branch of artificial intelligence.It briefly describes the current research status on the use of reinforcement learning techniques,addresses the difficulties in unmanned aerial vehicles networking,and outlines the main ideas for ap-plying reinforcement learning techniques in this field by combining them with the characteristics of unmanned aerial vehicles networking.The paper reviews several application scenarios and key networking technologies,highlighting the opportunities and challenges encountered by intelligent unmanned aerial vehicles networking technology based on reinforcement learning.The paper concludes that the research enhances unmanned aerial vehicles communica-tion by improving perception and decision-making capabilities,meeting the needs of dynamically changing environ-ments that demand a high degree of autonomy.Additionally,it provides valuable theoretical foundations and practi-cal guidance for the future development of intelligent unmanned aerial vehicles networking technology.
关键词
飞行自组网/强化学习/深度Q网络算法/多智能体/无人机集群/智能路由/资源分配/跨层优化
Key words
flying Ad Hoc network/reinforcement learning/deep Q-network algorithm/multiagent/unmanned aeri-al vehicles swarm/intelligent routing/resource allocation/cross-layer optimization
维普
万方数据