认知无线传感器网络功率控制研究

Study on power control of cognitive wireless sensor networks

为了提高无线传感器网络对频谱资源的利用率,文章在无线传感器网络中引入了认知无线电,利用了深度强化学习中提出的完全去中心化MARL算法实现了对无线传感器网络功率控制,算法是分布式的,非常适合大规模无线传感器网络。缺点是引入了认知无线电、多智能体深度强化学习会使节点成本上升。

In order to improve the utilization of spectrum resources in wireless sensor networks,we introduce cognitive radio in wireless sensor networks.The completely decentralized MARL algorithm proposed in depth reinforcement learning is used to realize the power control of wireless sensor networks.The algorithm is distributed and is very suitable for large-scale wireless sensor networks.The disadvantage is that the introduction of cognitive radio,multi-agent deep reinforcement learning will increase the cost of nodes.