[論文レビュー] Reconfigurable Intelligent Surfaces-assisted Positioning in Integrated Sensing and Communication Systems

Huyen-Trang Ta, Ngoc-Son Duong+3 more|arXiv (Cornell University)|Feb 16, 2026

Advanced Wireless Communication Technologies被引用数 0

ひとこと要約

要約: The paper proposes a RIS-assisted ISAC positioning framework with a coarse parameter estimation stage followed by a fast RIS-aware refinement that decouples linear gains from nonlinear geometry via a modified Levenberg method, achieving comparable accuracy to traditional methods with reduced complexity.

ABSTRACT

This paper investigates the problem of high-precision target localization in integrated sensing and communication (ISAC) systems, where the target is sensed via both a direct path and a reconfigurable intelligent surface (RIS)-assisted reflection path. We first develop a sequential matched-filter estimator to acquire coarse angular parameters, followed by a range recovery process based on subcarrier phase differences. Subsequently, we formulate the target localization problem as a non-linear least squares optimization, using the coarse estimates to initialize the target's position coordinates. To solve this efficiently, we introduce a fast iterative refinement algorithm tailored for RIS-aided ISAC environments. Recognizing that the signal model involves both linear path gains and non-linear geometric dependencies, we exploit the separable least-squares structure to decouple these parameters. Furthermore, we propose a modified Levenberg algorithm with an approximation strategy, which enables low-cost parameter updates without necessitating repeated evaluations of the full non-linear model. Simulation results show that the proposed refinement method achieves accuracy comparable to conventional approaches, while significantly reducing algorithmic complexity.

研究の動機と目的

• Motivate high-precision target localization in RIS-enabled ISAC systems.
• Develop a dual-path sensing model combining direct and RIS-reflected echoes to enhance localization accuracy.
• Propose a two-stage estimation framework (coarse estimation plus fast iterative refinement) to exploit the separable structure of the problem.
• Introduce a low-cost Levenberg-based refinement that updates parameters efficiently by leveraging linear-gain/nonlinear-geometry decoupling.

提案手法

• Model a 2D RIS-assisted ISAC system with direct and RIS paths and derive a nonlinear least-squares (NLS) formulation for target localization.
• Construct coarse estimation using a direct-path dictionary and RIS-path dictionary to obtain initial angular and range estimates.
• Formulate a two-path joint estimation problem where delays and angles are tied to the target position via geometry.
• Develop a fast refinement algorithm that uses a separable least-squares approach to decouple linear path gains from nonlinear geometric parameters.
• Use a modified Levenberg update with Taylor-linearized atoms to update the target position with reduced computational burden.

実験結果

リサーチクエスチョン

• RQ1 Can BIS/ RIS-assisted ISAC achieve accurate target localization using dual echo paths (direct and RIS-reflected) compared to single-path models?
• RQ2 How can coarse angular/range estimates be refined efficiently to reach high localization accuracy in RIS-aided ISAC systems?
• RQ3 Does exploiting the separable structure (linear gains vs nonlinear geometry) enable a low-cost optimization without sacrificing accuracy?
• RQ4 What is the computational trade-off between conventional CD+GD refinement and the proposed linearization-based refinement in RIS-enabled ISAC?
• RQ5 How does the proposed method perform under varied dictionary sizes and SNR conditions?

主な発見

• A two-stage estimation strategy achieves high localization precision, with coarse estimates followed by a fast RIS-aware refinement.
• The refinement decouples linear channel gains from nonlinear geometry, enabling a lower-cost update via a small 2x2 Levenberg step.
• Simulation results show the proposed refinement attains accuracy comparable to conventional methods while reducing algorithmic complexity.
• Coarse estimation benefits from a larger angular dictionary, improving RMSE for both direct and RIS paths, with RIS performance limited by lower effective SNR and error propagation from the direct path.
• The refined method converges to similar noise-limited performance as CD+GD at moderate-to-high SNR, while exhibiting slightly different convergence behavior at very low SNR.
• The complexity analysis indicates the proposed method can achieve speedups proportional to the inner update count due to reduced expensive atom/Jacobian reconstructions.