基于知识图谱和Softmax回归的干扰信号识别方法

Recognition Method of Interference Signal Based on Knowledge Graph and Softmax Regression

随着信息技术的发展,未来移动通信系统将是地海空一体化的综合性网络,因此通信系统所面临的干扰情况将异常复杂,这将导致干扰信号的识别与抑制任务更具有挑战性。针对此问题,文中提出一种基于知识图谱和Softmax回归的干扰信号识别方法。该方法首先构建干扰信号识别的专家知识图谱并利用TransR算法将其嵌入到低维向量空间中;然后,提取出每个干扰信号的多维特征并做归一化处理,将这些归一化特征值作为Softmax回归模型的输入,从而提高Softmax回归算法对干扰信号识别的准确率。仿真结果表明,对于典型的干扰样式,文中方法比基于Softmax回归和基于BP神经网络的干扰信号识别方法具有更好的识别性能。

With the development of information technology,the future mobile communication system will be a comprehensive network integrating land,sea and air. Therefore,the interference situation faced by the communication system will be extremely complex,which will make the task of identification and suppression of interference signals more challenging. To solve this problem,an interference signal recognition method based on knowledge graph and Softmax regression is proposed in this paper. The expert knowledge map of interference signal recognition is firstly constructed and embedded into the low-dimensional vector space by using the TransR algorithm. Then,the multi-dimensional features of each interference signal are extracted and normalized. These normalized eigenvalues are taken as the input of the Softmax regression model. Thus,the accuracy of Softmax regression algorithm for interference signal identification is improved. Simulation results show that for typical interference patterns,the proposed method has better performance than the methods based on Softmax regression and BP neural network.