Integrated Sensing, Identification, and Backscatter Communications: A Tensor-Based Multi-Antenna Framework

Abstract

Existing integrated sensing and communications (ISAC) systems are generally confined to sensing physical states of targets (e.g., positions and velocities), yet remain incapable of acquiring target identities. In light of this, we incorporate a backscatter communications (BackCom)-based target identification system into a frequency-modulated continuous wave (FMCW)-enabled multi-antenna sensing system, forming an integrated sensing, identification, and backscatter communications (ISIBC) system. In this paper, our main focus is to achieve three-dimensional (3D) localization and radial velocity estimation of the targets, and to simultaneously detect the transmitted backscatter device (BD) symbols from the FMCW echo signals, thereby acquiring the target identity information. To accomplish the above tasks, we propose a tensor-based processing framework. Specifically, the discrete beat signal received by a uniform planar array (UPA) is first formulated as an explicit fourth-order tensor model. Exploiting the property that the factor matrices of the fourth-order tensor adhere to the Vandermonde structure, we propose a novel non-iterative tensor decomposition algorithm. Then, utilizing the amplitude characteristics of the digital sinc function, we propose a computationally efficient scheme to jointly estimate the targets’ azimuth angles, elevation angles, and ranges for 3D localization, while simultaneously estimating their radial velocities and detecting the BD symbols from the estimated factor matrices. Finally, extensive simulation results are presented to demonstrate the effectiveness and superior performance of the tensor-based joint 3D localization, radial velocity estimation, and identification framework.