RIS-ASSISTED NEAR-FIELD ISAC FOR MULTI-TARGET INDICATION IN NLOS SCENARIOS
Hang Ruan, Homa Nikbakht, Ruizhi Zhang, Honglei Chen, Yonina C. Eldar · 2025
Read RIS-ASSISTED NEAR-FIELD ISAC FOR MULTI-TARGET INDICATION IN NLOS SCENARIOS by Hang Ruan, Homa Nikbakht, Ruizhi Zhang, Honglei Chen, Yonina C. Eldar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-08642.txt.
Introduction
Hang Ruan and 4 co-authors propose a beamforming method for sensing several targets without a direct line of sight in a 2025 paper. Their system combines sensing and communication, known as ISAC, in the near field. A reconfigurable intelligent surface, a panel that steers signals, creates a path around blocks.
Classic measures of beam strength and target separation are extended to use both angle and distance. The design jointly sets base station beams and surface phase shifts while keeping communication rates. An alternating optimization method solves the hard problem step by step. Simulations show clear sensing of targets that share an angle but differ in distance.
Catalog record
| CREATOR | Hang Ruan, Homa Nikbakht, Ruizhi Zhang, Honglei Chen, Yonina C. Eldar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-08642
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
RIS-ASSISTED NEAR-FIELD ISAC FOR MULTI-TARGET INDICATION IN NLOS SCENARIOS Hang Ruan, Homa Nikbakht, Ruizhi Zhang, Honglei Chen, Yonina C. Eldar Abstract Enabling multi-target sensing in near-field integrated sensing and communication (ISAC) systems is a key challenge, particularly when line-of-sight paths are blocked. This paper proposes a beamforming framework that leverages a reconfigurable intelligent surface (RIS) to achieve multi-target indication. Our contribution is the extension of classic beampattern gain and inter-target cross-correlation metrics to the near-field, leveraging both angle and distance information to discriminate between multiple users and targets. We formulate a problem to maximize the worst-case sensing performance by jointly designing the beamforming at the base station and the phase shifts at the RIS, while guaranteeing communication rates. The non-convex problem is solved via an efficient alternating optimization (AO) algorithm that utilizes semidefinite relaxation (SDR). Simulations demonstrate that our RIS-assisted framework enables high-resolution sensing of co-angle targets in blocked scenarios. † † address: 1 Weizmann Institute of Science, 2 Princeton University 3 University of Electronic Science and Technology of China, 4 MathWorks, Inc. 1 Introduction Future 6G networks require high-accuracy sensing alongside high-speed communication, for which integrated sensing and communication (ISAC) is a pivotal technology [ 1 ] . By sharing spectrum and hardware, ISAC improves spectral efficiency, reduces costs, and creates synergy between the two functions [ 1 ] . The trend towards large antenna arrays and high frequency pushes ISAC into the near-field regime, where the spherical wavefront provides a distance dimension [ 2 , 3 ] . This critically enables the discrimination of co-angle targets, which is impossible in the far-field [ 2 , 3 ] . Concurrently, reconfigurable intelligent surfaces (RISs) enhance ISAC systems by creating virtual links to overcome signal blockages [ 4 ] . While this has spurred research into RIS -assisted near-field ISAC [ 5 , 6 , 7 , 8 , 9 ] , these studies focus on single-target scenarios and cannot sense multiple targets simultaneously. Recognizing this limitation, a growing body of research has started to address the more challenging multi-user, multi-target ISAC problem with RIS assistance. These works have explored various performance metrics to manage the intricate trade-offs, such as the sensing signal-to-interference-plus-noise ratio (SINR) [ 10 ] , beam pattern gain [ 11 ] , mutual information (MI) [ 12 ] , and detection probability [ 13 , 14 ] . However, these works are fundamentally limited by their far-field assumptions, as their angle-only sensing metrics are insufficient for near-field scenarios where resolving targets in the same direction requires exploiting the distance dimension. To effectively harness near-field propagation for multi-target sensing, our prior work in [ 15 ] extended the robust metrics of transmit beampattern gain and inter-target cross-correlation to the near-field. This approach is offers two distinct advantages. First, unlike metrics such as sensing SINR , MI , or detection probability, our approach is agnostic to the targets’ often unknown radar cross-sections (RCSs) . Second, the explicit suppression of inter-target cross-correlation is critical for accurate multi-target indication; without it, metrics based on SINR , beampattern gain only or detection probability are insufficient for resolving multiple targets with classical multiple-input-multiple-output (MIMO) techniques [ 16 , 17 ] . However, the framework in [ 15 ] was limited to line-of-sight (LoS) -only scenarios. This paper bridges that critical gap by introducing a RIS to overcome non-line-of-sight (NLoS) blockages in a multi-user, multi-target near-field ISAC system. Our proposed framework leverages the near-field’s spherical wavefront to resolve targets in both angle and distance, even when obstructed. The main contributions of this work are as follows: 1) We formulate a joint transmission beamforming and RIS response optimization problem for RIS -assisted near-field ISAC that accommodates multiple NLoS users and targets. 2) To enable sensing of multiple targets, we first extend classical far-field beam-pattern and cross-correlation metrics to the near-field, enabling the discrimination of objects at identical directions but different distances. 3) We develop an efficient alternating optimization (AO) algorithm to solve the complex non-convex problem. The sub-optimizations of both transmission beamforming and RIS response are solved via the semidefinite relaxation (SDR) technique. 4) We demonstrate through numerical simulations that our proposed RIS-assisted design overcomes blockages to achieve significant performance gains, enabling high-resolution sensing and communication in scenarios. 2 System and Signal Model 2.1 System Model We consider a near-field mmWave ISAC system (Fig. 1 ) comprising a base station (BS) , an RIS , K single-antenna users in the set \mathcal{K} , and L point-targets in the set \mathcal{L} . The BS is equipped with an N_{t} -element transmit array and an N_{r} -element receive array, with antenna positions \bm{p}_{t}(n_{t}),\bm{p}_{r}(n_{r})\in\mathbb{R}^{3} and corresponding normal vectors \bm{w}_{t},\bm{w}_{r} defining the main radiation directions. The RIS contains N_{s} passive elements, with the n_{s} -th element’s position and normal vector being \bm{p}_{s}(n_{s}) and \bm{w}_{s} , respectively. The position of any user or target a\in\mathcal{K}\cup\mathcal{L} is \bm{p}_{a} . To model potential blockages, an indicator \alpha_{a}\in\{0,1\} describes the LoS status ( \alpha_{a}=1 for an available path, 0 otherwise). All entities operate within the BS ’s near-field region, bounded by the Rayleigh distance R_{\text{RL}}=2D^{2}/\lambda , where D is the BS antenna aperture and \lambda is the signal wavelength. Fig. 1: Illustration of the RIS-assisted near-field ISAC. 2.2 Transmitted ISAC Signal Under a fully-digital beamforming architecture [ 18 ] , the signal transmitted from the BS at time t ( t=1,\dots,T , where T is the block length) is given by \bm{x}[t]=\sum_{k\in\mathcal{K}}\bm{f}_{k}c_{k}[t]+\bm{s}[t] , where \bm{f}_{k}\in\mathbb{C}^{N_{t}} is the communication beamforming vector for user k , and c_{k}[t] is the corresponding data symbol, satisfying \mathbb{E}\left\{c_{k}[t]c_{l}^{*}[t]\right\}=\delta[k-l] . The term \bm{s}[t]\in\mathbb{C}^{N_{t}} represents the dedicated sensing signal with covariance \mathbf{R}_{s}=\mathbb{E}\left\{\bm{s}[t]\bm{s}^{H}[t]\right\} . The total transmit covariance is: \mathbf{R}_{x}=\mathbb{E}\left\{\bm{x}[t]\bm{x}^{H}[t]\right\}=\sum\nolimits_{k\in\mathcal{K}}\bm{f}_{k}\bm{f}_{k}^{H}+\mathbf{R}_{s}. (1) 2.3 Communication Model For user k\in\mathcal{K} , the overall downlink channel \bm{h}_{t,k}(\mathbf{\Phi})\in\mathbb{C}^{N_{t}} is a sum of the LoS path (if it exists) and the BS - RIS -user path: \bm{h}_{t,k}(\mathbf{\Phi})={\alpha_{k}\bar{\bm{h}}_{t,k}}+{\left(\bm{h}_{s,k}^{T}\mathbf{\Phi}\mathbf{G}_{t}\right)^{T}}. (2) Following the near-field channel model [ 19 ] , the components of the LoS channel, \bar{\bm{h}}_{t,k}\in\mathbb{C}^{N_{t}} , are defined as: \left[\bar{\bm{h}}_{t,k}\right]_{n_{t}}=\sqrt{F(\bm{p}_{k}-\bm{p}_{t}(n_{t}),\bm{w}_{t})}\beta(\bm{p}_{k}-\bm{p}_{t}(n_{t})), (3) where F(\bm{p}_{k}-\bm{p}_{t}(n_{t}),\bm{w}_{t}) is the radiation pattern, given by: F(\bm{p},\!\bm{w})\!=\!\begin{cases}2(b\!+\!1)\cos^{b}(\psi(\bm{p},\!\bm{w})),&\psi(\bm{p},\!\bm{w})\!\in\![0,\pi/2],\\ 0,&\text{otherwise},\end{cases} where \psi(\bm{p},\bm{w}) is the angle between the direction vector \bm{p} and the array’s normal vector \bm{w} . The parameter b determines the Boresight gain, which is 2 for the dipole. The term \beta(\bm{p}_{k}-\bm{p}_{t}(i,l)) denotes the attenuation and phase change caused by the path, with \beta(\bm{p}) being \beta(\bm{p})={\lambda}/\left({4\pi\left\lVert\bm{p}\right\rVert}\right)\cdot e^{-j\frac{2\pi}{\lambda}\left\lVert\bm{p}\right\rVert}. (4) The BS - RIS -user path consists of three components: the BS-RIS channel \mathbf{G}_{t}\in\mathbb{C}^{N_{s}\times N_{t}} , the RIS reflection matrix \mathbf{\Phi}\in\mathbb{C}^{N_{s}\times N_{s}} , and the RIS-user channel \bm{h}_{s,k}\in\mathbb{C}^{N_{s}} . Their elements are: \displaystyle\begin{split}[\mathbf{G}_{t}]_{n_{s},n_{t}}\!=&\beta(\bm{p}_{s}\!(n_{s})\!-\!\bm{p}_{t}(n_{t}))\!\sqrt{\!F\!(\bm{p}_{s}(n_{s})\!-\!\bm{p}_{t}(n_{t}),\!\bm{w}_{t})}\\ &\times\sqrt{F(\bm{p}_{t}(n_{t})\!-\!\bm{p}_{s}(n_{s}),\bm{w}_{s})},\end{split} (5) \displaystyle\mathbf{\Phi}= \displaystyle\mbox{diag}(e^{j\theta_{1}},\dots,e^{j\theta_{N_{s}}}), (6) \displaystyle\left[\bm{h}_{s,k}\right]_{n_{s}}= \displaystyle\sqrt{F(\bm{p}_{k}-\bm{p}_{s}(n_{s}),\bm{w}_{s})}\beta(\bm{p}_{k}-\bm{p}_{s}(n_{s})), (7) where \theta_{n_{s}}(n_{s}=1,\dots,N_{s}) is the response of the n_{s} -th RIS element. The received signal at user k is y_{k}[t]=\bm{h}_{t,k}^{T}(\mathbf{\Phi})\bm{x}[t]+n_{k}[t], (8) where n_{k}[t]\sim\mathcal{CN}(0,\sigma_{k}^{2}) is the additive noise. Then, the SINR is: \eta_{k}=\frac{\left\lvert\bm{h}_{t,k}^{T}(\mathbf{\Phi})\bm{f}_{k}\right\rvert^{2}}{\sum_{j\neq k}\left\lvert\bm{h}_{t,k}^{T}(\mathbf{\Phi})\bm{f}_{j}\right\rvert^{2}+\bm{h}_{t,k}^{T}(\mathbf{\Phi})\mathbf{R}_{s}\bm{h}_{t,k}^{*}(\mathbf{\Phi})+\sigma_{k}^{2}}. (9) The achievable communication rate for user k is used as the performance metric, given by [ 19 ] : R_{k}=\log_{2}(1+\eta_{k}). (10) 2.4 Sensing Model The sensing channel constitutes a round-trip path, consisting of the forward path from the BS to a target and the backward path from the target to the BS. Similarly to ( 2 ), the forward and backward channels, denoted as \bm{h}_{t,l}(\mathbf{\Phi})\in\mathbb{C}^{N_{t}} and \bm{h}_{l,r}(\mathbf{\Phi})\in\mathbb{C}^{N_{r}} , are composed of the direct