Mobile communication networks were designed to mainly support ubiquitous wireless communications, yet they are also expected to achieve radio sensing capabilities in the near future. However, most prior studies on radio sensing usually rely on far-field assumption with uniform plane wave (UPW) models. With the ever-increasing antenna size, together with the growing demands to sense nearby targets, the conventional far-field UPW assumption may become invalid. Therefore, this paper studies near-field radio sensing with extremely large-scale (XL) antenna arrays, where the more general uniform spheric wave (USW) sensing model is considered. Closed-form expressions of the Cram\'er-Rao Bounds (CRBs) for both angle and range estimations are derived for near-field XL-MIMO radar mode and XL-phased array radar mode, respectively. Our results reveal that different from the conventional UPW model where the CRB for angle decreases unboundedly as the number of antennas increases, for XL-MIMO radar-based near-field sensing, the CRB decreases with diminishing return and approaches to a certain limit as the number of antennas increases. Besides, different from the far-field model where the CRB for range is infinity since it has no range estimation capability, that for the near-field case is finite. Furthermore, it is revealed that the commonly used spherical wave model based on second-order Taylor approximation is insufficient for near-field CRB analysis. Extensive simulation results are provided to validate our derived CRBs.

翻译：移动通信网络最初设计主要用于支持普适无线通信，但未来有望实现无线电感知能力。然而，现有大多数无线电感知研究通常基于远场假设，采用均匀平面波（UPW）模型。随着天线尺寸日益增大，以及对近距目标感知需求的增长，传统远场UPW假设可能不再适用。为此，本文研究了基于极大规模天线阵列的近场无线电感知，采用更具普适性的均匀球面波（USW）感知模型。针对近场XL-MIMO雷达模式和XL相控阵雷达模式，分别推导了角度与距离估计的克拉美-罗界（CRB）闭合表达式。研究结果表明：与常规UPW模型中角度CRB随天线数量增加无界下降不同，基于XL-MIMO雷达的近场感知中，CRB随天线数量增加的递减效应逐渐减弱，并趋近于某一极限值。此外，与远场模型中距离CRB为无穷大（即无距离估计能力）不同，近场场景下距离CRB为有限值。分析进一步揭示，基于二阶泰勒近似的常见球面波模型不足以支撑近场CRB分析。通过大量仿真验证了所推导CRB的有效性。