基于压缩感知的辐射源信号数据级融合识别方法

发布时间：2018-05-18 02:38

本文选题：兵器科学与技术 + 辐射源识别　；参考：《兵工学报》2017年08期

【摘要】：针对协同侦察数据级融合识别中通信量大的问题,利用压缩感知可用少量测量值表征完整信号的特点,提出了一种基于压缩感知的数据级融合识别方法。接收终端采用确定测量阵对侦察信号的Gabor时频特征进行压缩测量,通过传输少量的压缩测量值以减轻通信压力,融合中心根据多源测量数据间的相关特性,采取相关性融合规则直接对多源压缩测量数据进行融合,最后再计算融合压缩测量值在不同类信号字典下的重构误差,最小重构误差对应的信号类别即识别结果。分别对融合识别方法识别性能和相关性融合规则融合效果进行仿真分析,实验结果表明:所提方法在保证识别率的同时大幅减小了数据通信代价,在低信噪比时识别性能突出、抗噪声干扰性能好;相比于其他融合规则,基于测量向量相关性的融合规则可保留更为全面的信息。
[Abstract]:Aiming at the problem of large traffic in data level fusion recognition of cooperative reconnaissance, a data level fusion recognition method based on compressed sensing is proposed, in which a small number of measurements can be used to represent the complete signal. The receiving terminal uses a certain measurement array to compress and measure the Gabor time-frequency characteristics of the reconnaissance signal. By transmitting a small amount of compressed measurement values to reduce the communication pressure, the fusion center is based on the correlation characteristics between the multi-source measurement data. The correlation fusion rule is adopted to fuse the multi-source compression measurement data directly. Finally, the reconstruction error of the fusion compression measurement value in different kinds of signal dictionaries is calculated. The category of the signal corresponding to the minimum reconstruction error is the recognition result. The performance of fusion recognition method and the fusion effect of correlation fusion rules are simulated and analyzed respectively. The experimental results show that the proposed method can greatly reduce the cost of data communication while ensuring the recognition rate, and the performance of the proposed method is outstanding when the signal-to-noise ratio is low. Compared with other fusion rules, the fusion rules based on measurement vector correlation can retain more comprehensive information.
【作者单位】：空军工程大学航空航天工程学院;
【基金】：航空科学基金项目(20152096019)
【分类号】：TN971

【相似文献】