Yang Yang, Yang Zhao, Haibo Wang, Dong Cao, Linyan Liu
{"title":"基于图注意网络的脉冲多普勒雷达目标检测方法","authors":"Yang Yang, Yang Zhao, Haibo Wang, Dong Cao, Linyan Liu","doi":"10.1109/icet55676.2022.9824889","DOIUrl":null,"url":null,"abstract":"Target detection under strong clutter is a relevant issue in many civilian and military application scenarios. Structured data based Deep Neural Networks (DNNs), such as Convolutional Neural Networks (CNNs), can automate feature extraction of radar time series and achieve better detection performance. We present in this work a target detection method to exploit spatial-temporal joint information of signal samples between Pulse Doppler (PD) radar range cells. The key point of proposed solution relies on the transformation of traditional signal time-series to graph-structured data, and application of Graph Attention Networks (GAT) to classify graph-structured data nodes. The presented results, achieved by processing PD radar data, demonstrate the validity of proposed method.","PeriodicalId":166358,"journal":{"name":"2022 IEEE 5th International Conference on Electronics Technology (ICET)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"The Target Detection Method of Pulse-Doppler Radar Based on Graph Attention Networks\",\"authors\":\"Yang Yang, Yang Zhao, Haibo Wang, Dong Cao, Linyan Liu\",\"doi\":\"10.1109/icet55676.2022.9824889\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Target detection under strong clutter is a relevant issue in many civilian and military application scenarios. Structured data based Deep Neural Networks (DNNs), such as Convolutional Neural Networks (CNNs), can automate feature extraction of radar time series and achieve better detection performance. We present in this work a target detection method to exploit spatial-temporal joint information of signal samples between Pulse Doppler (PD) radar range cells. The key point of proposed solution relies on the transformation of traditional signal time-series to graph-structured data, and application of Graph Attention Networks (GAT) to classify graph-structured data nodes. The presented results, achieved by processing PD radar data, demonstrate the validity of proposed method.\",\"PeriodicalId\":166358,\"journal\":{\"name\":\"2022 IEEE 5th International Conference on Electronics Technology (ICET)\",\"volume\":\"10 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-05-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2022 IEEE 5th International Conference on Electronics Technology (ICET)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/icet55676.2022.9824889\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 IEEE 5th International Conference on Electronics Technology (ICET)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/icet55676.2022.9824889","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
The Target Detection Method of Pulse-Doppler Radar Based on Graph Attention Networks
Target detection under strong clutter is a relevant issue in many civilian and military application scenarios. Structured data based Deep Neural Networks (DNNs), such as Convolutional Neural Networks (CNNs), can automate feature extraction of radar time series and achieve better detection performance. We present in this work a target detection method to exploit spatial-temporal joint information of signal samples between Pulse Doppler (PD) radar range cells. The key point of proposed solution relies on the transformation of traditional signal time-series to graph-structured data, and application of Graph Attention Networks (GAT) to classify graph-structured data nodes. The presented results, achieved by processing PD radar data, demonstrate the validity of proposed method.