Reconfigurable Intelligent Surface-Assisted Localization in OFDM Systems With Carrier Frequency Offset and Phase Noise

Hanfu Zhang, Erwu Liu, Rui Wang, Wei Ni, Zhe Xing, Yan Liu, Abbas JamalipourIEEE Trans. Wireless Communications 2025 · DOI

TL;DR

This paper introduces a novel algorithm for highly accurate localization in OFDM systems using Reconfigurable Intelligent Surfaces (RIS), even in the presence of hardware imperfections like carrier frequency offset (CFO) and phase noise (PN). By jointly optimizing RIS phase shifts and estimating these impairments, the proposed method achieves localization accuracy improvements of up to two orders of magnitude.

What Problem Does This Paper Address?

Achieving centimeter-level localization accuracy is crucial for emerging technologies like the Industrial Internet of Things. While Reconfigurable Intelligent Surfaces (RIS) show great promise for enhancing wireless localization by creating virtual line-of-sight paths, most existing research assumes ideal hardware. However, real-world systems suffer from imperfections like Carrier Frequency Offset (CFO) and Phase Noise (PN) due to oscillator inaccuracies. These impairments can significantly degrade localization performance by causing signal distortions. This paper tackles the challenge of developing a localization system that accounts for both RIS assistance and these critical hardware imperfections, specifically within the context of OFDM systems.

Methods

The core of this paper is a novel Alternating Optimization (AO)-based iterative algorithm designed to jointly estimate the UE's position, CFO, and PN. Key methodological contributions include:

  • Closed-form expressions for CFO and PN estimation: For each iteration, given the UE's position, the algorithm derives closed-form solutions for CFO and PN, significantly reducing computational complexity and improving stability.
  • Novel RIS phase shift optimization: A new algorithm is proposed to optimize RIS phase shifts by minimizing the analytical lower bound of localization accuracy. This involves using semidefinite relaxation (SDR) and the Schur complement to transform a non-convex optimization problem into a solvable semidefinite program (SDP).
  • Hybrid Cramér-Rao Lower Bound (HCRLB) analysis: The paper derives the HCRLB for the joint estimation problem, providing a theoretical benchmark for performance and serving as an objective function for RIS phase shift optimization.
  • Approximation of PN vector: To manage complexity, the PN vector is approximated using a truncated Singular Value Decomposition (SVD) based on the redundancy observed in its covariance matrix.
  • Gradient Descent (GD) for position estimation: Given the CFO and PN estimates, a GD algorithm is employed to iteratively refine the UE's position.

Datasets

Not discussed in the paper. The paper relies on numerical simulations to evaluate its proposed algorithms.

Key Insights and Ideas

  • Hardware imperfections are critical: The study highlights that ignoring CFO and PN can lead to significant performance degradation in RIS-assisted localization, underscoring the need for algorithms that account for these real-world issues.
  • Joint optimization is key: Simultaneously estimating position, CFO, and PN, along with optimizing RIS phase shifts, yields substantial improvements in localization accuracy compared to methods that address these aspects in isolation or assume ideal conditions.
  • RIS phase shift optimization is crucial: The paper demonstrates that intelligently configuring RIS phase shifts, rather than using random configurations, can improve localization accuracy by orders of magnitude.
  • Near-field localization benefits from RIS: The research focuses on near-field scenarios, where RIS plays a vital role in overcoming signal blockage and enhancing localization precision.

Key Results

  • Significant accuracy improvement: The proposed algorithm achieves localization accuracy enhancements of up to two orders of magnitude compared to existing methods.
  • Near-analytical performance: The localization accuracy of the proposed algorithm is shown to be very close to the analytical lower bound, with a root mean square error (RMSE) lower than 10⁻² m.
  • RIS optimization impact: Optimized RIS phase shifts improve the Position Error Bound (PEB) by approximately two orders of magnitude compared to random RIS phase shifts and by more than threefold compared to methods that ignore hardware impairments.
  • Effect of transmit power and RIS size: Increasing transmit power and the number of RIS elements consistently improves localization accuracy.
  • Robustness to PN: Decreasing PN variance leads to significant improvements in the estimation performance of CFO, PN, and UE position.

Limitations and Future Work

  • Computational complexity: While the AO approach reduces complexity, the paper acknowledges that optimizing the RIS phase shifts for a large number of PEBs related to random positions can still be computationally intensive.
  • Future work: The authors suggest exploring optimization algorithms with reduced complexity for scenarios requiring the averaging of a large number of PEBs. This would be crucial for practical, real-time applications.

Summary updated

BibTeX
@article{zhang2025reconfigurable,
  author = {Hanfu Zhang and Erwu Liu and Rui Wang and Wei Ni and Zhe Xing and Yan Liu and Abbas Jamalipour},
  title = {{Reconfigurable Intelligent Surface-Assisted Localization in OFDM Systems With Carrier Frequency Offset and Phase Noise}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2025},
  doi = {10.1109/twc.2025.3557935},
}
Figure