首页|期刊导航|雷达科学与技术|基于时序RD图的机载雷达平台自运动估计

基于时序RD图的机载雷达平台自运动估计OA

Temporal RD Map Based Self-Motion Estimation for Airborne Radar Platform

中文摘要英文摘要

机载雷达平台的自运动状态高精度实时感知是实现精确探测的关键前提.脉冲多普勒雷达的距离-多普勒(RD)图谱中蕴含了平台运动(如水平距离、高度、速度等)与地物散射等复杂耦合信息.然而,传统信号处理方法难以在复杂地形与噪声干扰下实现参数解耦,且分辨率受限.深度学习方法为解耦RD图的内蕴信息提供了新路径,但现有研究大多集中在单帧RD图的静态分析,忽视了平台运动这一连续动作的时序信息.针对上述问题,本文选取了专为时序RD图参数反演任务设计的RDMFNet,并将其与ResNet、UNet等通用图像网络进行对比.同时,为克服真实数据稀缺的问题,本文还构建了一个高保真仿真流程,通过地形镶嵌与建筑物嵌入构建复杂场景,并在回波仿真中引入阴影遮蔽算法,结合噪声注入得到高保真时序RD图.使用时序堆叠、时序融合等训练策略的实验结果表明,所选用的RDMFNet的运动参数估计性能显著优于ResNet和UNet,除此之外,时序信息的引入,使模型的抗噪能力大大提高的同时,进一步增强了模型的估计精度,为精度高、鲁棒性强的自运动估计提供了有效方案.

The high-precision and real-time perception of the self-motion state of the airborne radar platform is es-sential for accurate target detection.The range-Doppler(RD)map of pulse-Doppler radar contains complex and nonlin-early coupled information related to platform motion parameters(such as horizontal distance,height,and velocity,etc)as well as ground scattering characteristics.However,traditional signal processing methods struggle to decouple these pa-rameters under complex terrain and noise interference,and their achievable estimation accuracy is fundamentally con-strained.Deep learning provides an alternative data-driven paradigm for exploiting the intrinsic information embedded in RD maps,yet existing studies mostly focus on static analysis of single-frame RD map,overlooking the sequential na-ture of the platform motion.To address these challenges,this paper adopts RDMFNet,which is a network specifically de-signed for temporal RD map parameter estimation,and compares its performance with general-purpose image based net-works such as ResNet and UNet.Meanwhile,to overcome the scarcity of real data,a high-fidelity simulation pipeline is developed.The pipeline constructs complex scenes through terrain mosaic and building embedding,introduces shadow masking in echo simulation,and injects additive noise to generate realistic temporal RD sequences.Experimental re-sults demonstrate that incorporating temporal information significantly improves the motion parameter estimation perfor-mance and noise robustness.The proposed approach provides an effective solution for high-precision and robust self-motion estimation of airborne radar platforms.

何屹廷;刘子源;南苏桐;张家炜;谷源涛

清华大学深圳国际研究生院,广东 深圳 518055清华大学电子工程系,北京 100084清华大学电子工程系,北京 100084清华大学电子工程系,北京 100084清华大学电子工程系,北京 100084

信息技术与安全科学

运动参数估计距离-多普勒图时间序列深度学习

motion parameter estimationrange-Doppler maptime seriesdeep learning

《雷达科学与技术》 2026 (4)

389-397,9

国家自然科学基金(U2230201)

10.3969/j.issn.1672-2337.2026.04.004

评论