基于特征融合的多节点调制识别方法

针对传感器网络节点的分布式结构,给出了基于特征融合的多节点联合调制识别方法。首先利用似然比推导了特征融合的联合识别准则;然后依据准则给出了多节点调制识别的仿真性能;最后和基于决策融合的多节点调制识别方法进行了比较。仿真结果表明,在没有过多增加网络通信负载的情况下,特征融合相比决策融合提高了识别性能。

Modulation classification algorithm of multinodes based on features fusion

This paper proposed a modulation recognition algorithm of multinodes based on features fusion for the distributed structure of wireless sensor network nodes. Firstly, derived combining classification criterion of features fusion by using likelihood ratio. Then, presented the simulation performances under the criterion. Finally, it proposed the performance comparation of features fusion and decisions fusion. In the not too much increase network communication load, the features fusion algorithm has a better performance.