ISAC中复合各向异性运动目标的分层相干成像
Hierarchical Coherent Imaging of Composite Anisotropic Moving Targets in ISAC
AI总结:
提出MOSAIC方法,利用分布式用户设备和单个ISAC基站,通过分层相干与非相干结合、多普勒预补偿,实现复合运动目标的高分辨率成像与速度估计,相比现有方法图像质量提升超50%。
摘要:
在集成感知与通信(ISAC)网络中,分布式设备可以通过利用相位相干信号处理协同生成周围环境的无线电图像。然而,现有成像方法并不适用于具有多个独立运动扩展部件的复合运动目标。这是由于简化的各向同性散射模型以及缺乏补偿每个部件不同多普勒频移的方法,导致图像散焦。我们提出了MOSAIC,这是首个针对复合运动目标的分层成像方法,使用分布式用户设备(UE)和单个ISAC基站(BS)。MOSAIC生成每个目标部件的高分辨率图像并估计其速度矢量。在选定的UE簇内进行相干成像,这些UE观测每个部件的局部各向同性散射,而簇特定图像在大角度范围内非相干组合以改善重建。为了减轻多普勒引起的散焦,在相干成像前对多普勒分量进行预补偿,将限制转化为解析多个目标部件的额外手段。这还通过关联UE间的多普勒频率实现了低复杂度的速度估计。仿真表明,与现有方法相比,在Wasserstein距离方面图像质量提升超过50%,速度估计精度达到分米/秒级。
英文摘要:
In Integrated Sensing and Communication (ISAC) networks, distributed devices can cooperate to produce radio images of the surrounding environment by exploiting phase-coherent signal processing. However, existing imaging methods are not well-suited for composite moving targets with multiple independently moving extended parts. This is due to simplistic isotropic scattering models and the lack of methods to compensate for distinct Doppler shifts from each component, which leads to image defocusing. We propose MOSAIC, the first hierarchical imaging method for composite moving targets using distributed User Equipments (UEs) and a single ISAC Base Station (BS). MOSAIC generates high-resolution images of each target part and estimates its velocity vector. Coherent imaging is performed within selected clusters of UEs observing a locally isotropic scattering from each part, while cluster-specific images are combined non-coherently across wide angles to improve the reconstruction. To mitigate Doppler-induced defocusing, Doppler components are pre-compensated before coherent imaging, turning a limitation into an additional means of resolving multiple target parts. This also enables low-complexity velocity estimation by associating Doppler frequencies across UEs. Simulations show a 35% to 45% improvement in image quality compared to existing methods, in terms of Wasserstein distance, and dm/s-level velocity estimation accuracy.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。