基于深度卷积神经网络的多节点协同干扰识别方法

Multi-node Cooperative Jamming Recognition Method Based on Deep Convolutional Neural Network

通信抗干扰是复杂电磁环境下无线通信生存能力的核心问题,而干扰识别是通信抗干扰的重要前置环节。目前的干扰识别大多采用单节点进行干扰识别,而单节点干扰识别容易受到信道环境影响导致性能下降。针对无线通信网络在低干信比条件下单节点干扰正确识别率较低等问题,提出了一种基于深度卷积神经网络的多节点协同干扰识别方法,并设计了基于中心判决和基于硬判决的两种干扰识别算法。仿真结果表明,采用多节点协同干扰识别方法能够显著提升无线通信网络在低干信比条件下的干扰正确识别率,且基于硬判决的方法较基于中心判决的方法有更好的性能。

Communication anti-jamming is the core issue of wireless communication survivability in complex electromagnetic environment,while jamming recognition is the significant pre-stage of the communication anti-jamming process.At present,single-node method is mostly used for jamming recognition and easy to be affected by the channel environment,resulting performance degradation.In order to solve the problem of low jamming correct recognition rate of single-node method in wireless communication networks under the condition of low jamming-to-signal ratio,a multi-node cooperative jamming recognition method which based on deep convolution neural network is proposed in this paper.Two jamming recognition algorithms based on central decision and hard decision are designed.Simulation results show that the multi-node cooperative jamming recognition method can significantly improve the correct jamming recognition rate of wireless communication networks under the condition of low jamming-to-signal ratio,and the method based on hard decision has better performance than the method based on central decision.