不完整数据集的MFR辐射源识别方法研究
|
摘要
该文提出一种基于随机森林的不完整数据集的多功能雷达(MFR)辐射源识别方法,该方法在MFR辐射源波形单元识别框架基础上,首先对参数缺失的先验知识集进行多重划分,得到多个不含缺失参数的样本子集,然后删减冗余子集并利用随机森林算法对各个子集构建弱分类器,最后根据弱分类器对识别结果贡献率的不同,进行权值设定,得到最终的识别模型。仿真实验证实了提出的MDRF-WA方法能够提高少量先验知识条件下波形单元识别的准确率和鲁棒性,降低计算成本。
For the multi-function radar(MFR) emitter identification with incomplete data, aiming at the problems of prior knowledge demand, low accuracy and poor robustness, which exist in the conventional identification methods, a method of waveform unit identification based on incomplete prior knowledge set is proposed. Firstly, based on the MFR waveform unit identification framework, the original prior knowledge with parameter missing is multiply divided, and a number of subsets of samples without parameter missing are obtained. Secondly, the redundant subsets are removed and a weak classifier is constructed for each subset by using the random forest algorithm. Finally, the weight is set according to the contribution rate of each weak classifier to the identification result, and the final identification model is obtained. Simulation results confirm the validity of the MDRF-WA waveform unit identification method proposed, moreover, MDRF-WA method can make full use of known prior knowledge, reduce the computational cost and improve the robustness and accuracy of the waveform unit identification under the condition of small training samples.
For the multi-function radar(MFR) emitter identification with incomplete data, aiming at the problems of prior knowledge demand, low accuracy and poor robustness, which exist in the conventional identification methods, a method of waveform unit identification based on incomplete prior knowledge set is proposed. Firstly, based on the MFR waveform unit identification framework, the original prior knowledge with parameter missing is multiply divided, and a number of subsets of samples without parameter missing are obtained. Secondly, the redundant subsets are removed and a weak classifier is constructed for each subset by using the random forest algorithm. Finally, the weight is set according to the contribution rate of each weak classifier to the identification result, and the final identification model is obtained. Simulation results confirm the validity of the MDRF-WA waveform unit identification method proposed, moreover, MDRF-WA method can make full use of known prior knowledge, reduce the computational cost and improve the robustness and accuracy of the waveform unit identification under the condition of small training samples.
引文
[1]张光义.相控阵雷达原理[M].北京:国防工业出版社,2009.ZHANG Guang-yi.Principles of phased array radar[M].Beijing:National Defense Industry Press,2009.
[2]WANG A,KRISHNAMURTHY V.Signal interpretation of multifunction radars:Modeling and statistical signal processing with stochastic context free Grammar[J].IEEETransactions on Signal Processing,2008,56(3):1106-1119.
[3]ARASARATNAM I,HAYKIN S,KIRUBARAJAN T,et al.Tracking the mode of operation of multi-function radars[C]//IEEE Conference on Rader.[S.l.]:IEEE,2006:10.1109/radar.2006.1631804.
[4]PING Y,CHANG Y F,ZHOU Y,et al.Fast and scalable support vector clustering for large-scale data analysis[J].Knowledge and Information Systems,2015,43(2):281-310.
[5]JIA W,ZHAO D,SHEN T,et al.An optimized classification algorithm by BP neural network based on PLS and HCA[J].Applied Intelligence,2015,43(1):1-16.
[6]陈昌孝,何明浩,徐璟,等.雷达辐射源识别技术研究进展[J].空军预警学院学报,2014(1):1-5.CHEN Chang-xiao,HE Ming-hao,XU Jing,et al.Progress of study on recognition technology of radar emitter[J].Journal of Air Force Early Warning Academy,2014(1):1-5.
[7]陈维高,贾鑫,朱卫纲,等.基于机动特征辅助的MFR状态预测方法[J].电子学报,2018,46(6):1404-1409.CHEN Wei-gao,JIA Xin,ZHU Wei-gang,et al.MFR state prediction method based on aircraft maneuvering features assistance[J].Chinese Journal of Electronics,2018,46(6):1404-1409.
[8]陈维高,贾鑫,朱卫纲,等.基于HMM的雷达状态转移估计方法[J].北京航空航天大学学报,2017,43(10):2171-2180.CHEN Wei-gao,JIA Xin,ZHU Wei-gang,et al.The radar state transfer estimation algorithm based on HMM model[J].Journal of Beijing University of Aeronautics and Astronautics,2017,43(10):2171-2180.
[9]刘海军.雷达辐射源识别关键技术研究[D].长沙:国防科学技术大学,2010.LIU Hai-jun.Researches on identification key technology for radar emitter[D].Changsha:National University of Defense Technology,2010.
[10]严远亭.不完整数据集的多视角集成分类研究[D].合肥:安徽大学,2016.YAN Yuan-ting.Research on multi-view ensemble classification for incomplete data[D].Hefei:AnHui University,2016.
[11]胡峻峰.基于机器视觉的实木地板分选技术研究[D].哈尔滨:东北林业大学,2015.HU Jun-feng.Research on technology of Solid wood floor sorting based on machine vision[D].Harbin:Northeast Forestry University,2015.
[12]张葛祥.雷达辐射源信号智能识别方法研究[D].成都:西南交通大学,2005.ZHANG Ge-xiang.Intelligent recognition methods for radar emitter signals[D].Chengdu:Southwest Jiaotong University,2005.
[13]马爽.多功能雷达电子情报信号处理关键技术研究[D].长沙:国防科学技术大学,2013.MA Shuang.Research on ELINT signal processing key technologies for multifunction radar[D].Changsha:National University of Defense Technology,2013.
[2]WANG A,KRISHNAMURTHY V.Signal interpretation of multifunction radars:Modeling and statistical signal processing with stochastic context free Grammar[J].IEEETransactions on Signal Processing,2008,56(3):1106-1119.
[3]ARASARATNAM I,HAYKIN S,KIRUBARAJAN T,et al.Tracking the mode of operation of multi-function radars[C]//IEEE Conference on Rader.[S.l.]:IEEE,2006:10.1109/radar.2006.1631804.
[4]PING Y,CHANG Y F,ZHOU Y,et al.Fast and scalable support vector clustering for large-scale data analysis[J].Knowledge and Information Systems,2015,43(2):281-310.
[5]JIA W,ZHAO D,SHEN T,et al.An optimized classification algorithm by BP neural network based on PLS and HCA[J].Applied Intelligence,2015,43(1):1-16.
[6]陈昌孝,何明浩,徐璟,等.雷达辐射源识别技术研究进展[J].空军预警学院学报,2014(1):1-5.CHEN Chang-xiao,HE Ming-hao,XU Jing,et al.Progress of study on recognition technology of radar emitter[J].Journal of Air Force Early Warning Academy,2014(1):1-5.
[7]陈维高,贾鑫,朱卫纲,等.基于机动特征辅助的MFR状态预测方法[J].电子学报,2018,46(6):1404-1409.CHEN Wei-gao,JIA Xin,ZHU Wei-gang,et al.MFR state prediction method based on aircraft maneuvering features assistance[J].Chinese Journal of Electronics,2018,46(6):1404-1409.
[8]陈维高,贾鑫,朱卫纲,等.基于HMM的雷达状态转移估计方法[J].北京航空航天大学学报,2017,43(10):2171-2180.CHEN Wei-gao,JIA Xin,ZHU Wei-gang,et al.The radar state transfer estimation algorithm based on HMM model[J].Journal of Beijing University of Aeronautics and Astronautics,2017,43(10):2171-2180.
[9]刘海军.雷达辐射源识别关键技术研究[D].长沙:国防科学技术大学,2010.LIU Hai-jun.Researches on identification key technology for radar emitter[D].Changsha:National University of Defense Technology,2010.
[10]严远亭.不完整数据集的多视角集成分类研究[D].合肥:安徽大学,2016.YAN Yuan-ting.Research on multi-view ensemble classification for incomplete data[D].Hefei:AnHui University,2016.
[11]胡峻峰.基于机器视觉的实木地板分选技术研究[D].哈尔滨:东北林业大学,2015.HU Jun-feng.Research on technology of Solid wood floor sorting based on machine vision[D].Harbin:Northeast Forestry University,2015.
[12]张葛祥.雷达辐射源信号智能识别方法研究[D].成都:西南交通大学,2005.ZHANG Ge-xiang.Intelligent recognition methods for radar emitter signals[D].Chengdu:Southwest Jiaotong University,2005.
[13]马爽.多功能雷达电子情报信号处理关键技术研究[D].长沙:国防科学技术大学,2013.MA Shuang.Research on ELINT signal processing key technologies for multifunction radar[D].Changsha:National University of Defense Technology,2013.